Can one 'wear out' an editor with bad papers?

Funny question - should the fact that a journal's editor is 'human' be a factor in submitting manuscripts?

You don't have to take into account the journal editor as human being -- e.g. there is nothing wrong with submitting to an editor you have never heard of or met but whose credentials look relevant -- but in my experience (I am in mathematics) I get better service from editors with whom I have a prior acquaintance and a prior (positive!) professional relationship. So my answer to this would be If possible, yes, try to make this work in your favor.

Example - say I'm writing 4 manuscripts. The first three are 'bad' and the fourth is good. Then I submit them to the journal, one every 6 months or so. The first three will be rejected. But will they affect the ruling of my 4th submitted manuscript, even it is much better?

Like @Davidmh, I am wondering why you are submitting three manuscripts that you describe as 'bad'. (Admittedly you call them "'bad'", not "bad", but still...) I am also wondering why you are submitting manuscripts that you know will be rejected, but I gather here you are setting up the premise of a hypothetical. Finally, I wonder whether you have a good reason for submitting four manuscripts in such rapid succession to the same journal, whether they are accepted or not.

But let me answer your question first: yes, having three manuscripts rejected from the same journal can conceivably affect the way your fourth manuscript is received, even though in theory it should not, and you have already identified the reason: editorial boards consist of human beings. Three 'bad' manuscripts is enough to establish a pattern that you don't understand the value of your own work. People who are involved in processing your fourth paper will find it very difficult not to view it with that lens. Of course this does not at all mean that your paper will be rejected: if it's sufficiently good it will almost certainly be accepted by that journal or by some other journal of similar quality. (But that would happen no matter what...)

So I do not recommend the practice you propose. Just to address a few side issues:

1) Unless you have a clear reason to do so, publishing too often in the same journal also looks (moderately) bad.

2) Some journals turn out to have quotas on papers from the same author. Once I resubmitted to the same editor of the Proceedings of the AMS within a year after getting a paper accepted there, and he told me that he couldn't consider the paper. Another time I had a paper accepted at a leading combinatorics journal with a large backlog, and within a year of acceptance I submitted a followup paper (with a different set of authors) which went deeper on the same topic. The editorial board discussed the matter, decided that with their large backlog they didn't want to publish two papers on the same topic in such rapid succession...and then asked me whether I would be willing to combine the newly submitted paper with the already accepted paper! (Needless to say, the answer was no.) These stories show that editors of journals certainly do have memory of authors of past submissions, and even past successful submissions can be a reason to want to submit elsewhere in the short term.

3) Gian-Carlo Rota quipped that the mathematical community calculates the worth of a mathematician by taking their best paper and dividing by the total number of papers they've written. This is deliciously cruel and obviously not to be taken literally -- in fact one needs to show a certain level of research activity in many hiring and promotion decisions at most research universities -- but there is more than a grain of truth to it. At this point in my career I write about 3-4 papers per year. If I were (quite counterfactually, obviously) in a negotiation to trade X past publications for one publication in a journal which is one level higher than the best journal I've published in so far, then my starting offer for X would be more than 10. Rota is right that it could actually help to increase my research profile in certain elite circles by removing most of my publications that do not contain major breakthroughs. This is something to think about with regard to your "three 'bad', then one good" proposal.

Added: When I left this answer, I thought the OP was in the field of mathematics, based on his SE activity. Then I found a comment which says he is in computer science. So it would be useful to him to have an answer located in CS, as well. Some of what I say should carry over, but maybe not all.

Further Added: In the first version of this answer, I wrote "better service and also better results from editors with whom I have a prior acquaintance." After having thought about the italicized passage, it is not clear to me that it is true (so I took it out). For instance I can think of two journals that have rejected at least two and accepted none of my papers in recent years, and in each case I had previously met the editor and had positive professional relations with them. In fact when I get even one rejection from an editor I don't really know, all else being equal I will probably not resubmit to them: maybe they're not biased against me at all, but if I don't know them at all, why not just take my chances elsewhere? What I think is true is that an editor who knows me is much more likely to know where I am coming from professionally and (i) find me a competent and fair referee and (ii) expend some effort to make sure that the referee process is done in a reasonable amount of time. Also, when I talk about "knowing editors," of course I mean professionally: I'm not talking about spending time on their yacht. Someone whom I have met and talked to at a conference, whom I have done a good refereeing job and gotten thanked for makes a good choice for an editor to submit to. I am not a journal editor, but I do a lot of service-related academic tasks, and I will say: I never set out to do a bad job, but knowing that an especially good job would help out someone who has helped me out in the past can make me want to work a bit harder than otherwise. I imagine that many if not most people are similar: this is a positive form of social reinforcement.

Should you keep in mind that the editor is human? If I interpret your question literally then the answer is vacuously "Yes," since undoubtedly the editor is human, so if you want to achieve the best results then surely you would take that into account, along with any other information that can help you reach good decisions.

But I assume that what you really meant to ask was more along the lines of "To what extent should the fact that the editor is human play a factor in my submission decisions? Discuss."

Well, in that case I would say: to a very large extent. More generally, in academia reputation is really, really important. Your reputation is shaped over time by the quality of the products (papers, talks etc.) you deliver. A bad reputation from writing and submitting multiple poor papers will almost certainly result in a negative bias against you that could easily translate to a less kind and more judgmental treatment from editors and (in areas like mathematics that don't have double-blind reviews) reviewers.

It's worth mentioning that the effect also works in the other direction, and someone with a very good reputation after publishing a series of brilliant, groundbreaking papers could come up with a mediocre paper and still have a reasonable chance of being treated more kindly by the editor and reviewers. Is this unfair? Maybe -- or, one might argue, maybe not, if you consider that the editor and reviewers have limited time and need to invest them in the papers that they think are statistically most likely to advance science in a meaningful way. A paper from an author who has a good track record is statistically a safer bet. Of course, one should not get carried away with such reasoning and still give the paper you are reviewing a careful reading and do the best you can to stay impartial in your evaluation.

The moral of my answer is that there are many ways in which the three "bad" papers in the example could be turned into "average" or maybe even "good" papers by putting in additional work, and that it's very worth going to the trouble of doing this, because there is much more at stake than just those few papers - indeed, the quality of the three papers can have a significant effect on one's future submissions and career as well.

A steady stream of rejected submissions can definitely leave editors feeling annoyed, assuming the papers are definite rejects and not borderline or debatable cases. I rarely see this in practice, but when I do it leaves me speculating uncharitably about the authors. Are they so clueless that they have no idea how likely their papers are to be accepted, even given the clearly established pattern? Are they submitting as a protest, knowing that the papers will probably be rejected but feeling that we deserve to waste time rejecting them if we are going to be so unfair? Are they trying to game the system by submitting papers they know most referees will not recommend for publication, in the hope of slipping in by getting a careless or sympathetic referee?

I don't actively hold this against the authors, but I do try extra hard to make sure consistent standards are upheld and random fluctuations don't reward this submission strategy.

Of course I know these speculations are unfair, and I don't necessarily believe them about any particular case, but I have to wonder. What particularly annoys me is that conscientious, hard-working referees are a limited resource, which I'd rather not spend on scattershot submissions. I've never actually reached the point of explicitly suggesting to the authors that maybe they are submitting too many papers, but I've considered it.

Note that this might be field-dependent, and I'm in mathematics (where publishing practices differ somewhat from other fields, and where refereeing is a much more difficult task). I've sometimes heard people in other fields describe an iterative submission process where you always start at the top journals and move one step down with each rejection. If this is common in your field, then maybe editors are used to it.