CSS attribute selector does not work a href

Use the $ after your href. This will make the attribute value to match the end of the string.

a[href$='.pdf'] { /*css*/ }

JSFiddle: http://jsfiddle.net/UG9ud/

E[foo]        an E element with a "foo" attribute (CSS 2)
E[foo="bar"]  an E element whose "foo" attribute value is exactly equal to "bar" (CSS 2)
E[foo~="bar"] an E element whose "foo" attribute value is a list of whitespace-separated values, one of which is exactly equal to "bar" (CSS 2)
E[foo^="bar"] an E element whose "foo" attribute value begins exactly with the string "bar" (CSS 3)
E[foo$="bar"] an E element whose "foo" attribute value ends exactly with the string "bar" (CSS 3)
E[foo*="bar"] an E element whose "foo" attribute value contains the substring "bar" (CSS 3)
E[foo|="en"]  an E element whose "foo" attribute has a hyphen-separated list of values beginning (from the left) with "en" (CSS 2)

source: http://www.w3.org/TR/selectors/

The accepted answer (using a[href$='.pdf']) assumes that that a link to a pdf will always end with .pdf. That is not necessarily the case, as the link could have a query string or a hash fragment, for example with a UTM tracking code or a page number, in which case those links would not be matched. In fact depending on your application this could be the case for most links.

<a href="/manual.pdf?utm_source=homepage">A PDF File</a>
<a href="/manual.pdf#page=42">A PDF File</a>

If you want to ensure your rule is also applied in those cases you could match .pdf anywhere in the attribute using

a[href*='.pdf']

However this will then match some unlikely but unintended things, such as a subdomain our.pdf.domain.com/a-page. But we can narrow it down further, as we know we would only use it match pdfs that have a query string or hash fragment. If we combine the 3 cases we should match all pdf links.

a[href$='.pdf'], a[href*='.pdf?'], a[href*='.pdf#'] {
    background: red;
}