Is it fine to post my paper's preprint both on my blog and on arXiv?

If my paper is already on my blog, isn't it more likely to be flagged as plagiarized by any average plagiarism software?

This won't be a problem, unless you do something really weird, like post your paper on the blog under a different name. Lots of arXiv papers also appear elsewhere, and the arXiv couldn't function if this caused trouble.

There have been evidences of proper articles being rejected. This being the case, how does it ensure security?

I'm not sure I understand what you mean by security. Strictly speaking, nothing is ensured, but in practice I believe very few proper submissions are rejected (almost none). Unless you are an amateur researcher or are working on something unconventional, I don't think you need to worry about rejection by the arXiv.

The reason we post our papers first on arXiv is to prevent the content from being stolen by some anonymous person, who first rejects our paper upon submission and unlawfully uses the content elsewhere. How is arXiv different in this aspect?

It's not different in principle, but in practice I don't consider this a serious issue. I doubt any arXiv moderator has ever stolen a submission, since the risk/reward ratio would be terrible. (It makes little sense to try to steal a paper when the authors will presumably post it online somewhere else shortly, and when there's an electronic record that you rejected it from the arXiv.)

More generally, I disagree that this is a key reason to post to the arXiv, since theft by reviewers is rare. When people talk about establishing credit by posting to the arXiv, they generally aren't talking about preventing theft, but rather publicizing the work enough that nobody else will independently make the same discovery and claim credit.

According to the Arxiv itself, it applies some plagiarism checks but these are only applied to other articles within the Arxiv. Of course, there could still be some checks against publications elsewhere, but given that a considerable amount of papers is only published on the Arxiv after journal publication, these can only be a tool for moderators.

So if your paper will cause a warning to be thrown at all, this is only a warning for the moderator. Moreover, it would be a type of with a high false-positive rate, as it applies to any paper that was already published in a journal. Thus any reasonable reviewer will see that the paper on your blog is claiming the same authors and thus there is no plagiarism going on.

Note, however, that there are some journals out there that allow preprints to be published on the Arxiv but not on personal webpages or webpages with commercial interest (for which ads may suffice). I recommend checking Sherpa or your target journals’ copyright agreement before publishing anything.