This problem is seriously non-trivial because to do this, you need to have a lexicon that lists all the forms and says whether they're singular or plural. There are a lot of irregular plural forms in English, plus there are a lot of ambiguous singular forms that a simple "is there an 's' at the end?" algorithm would need to account for (like "bus").

Your best bet is to process the text you need the plural forms for via a POS (Part of Speech) Tagger like the Tree Tagger at the Department of Linguistic Processing at the University of Stuttgart. It's free as in beer, but you can't play with the code (which is C, so it runs very fast).

There are other taggers out there, including one packaged with a bunch of Perl tools called Xlex (not the remedy for constipation). It's written in C++ and you can play with it online. They will send you the whole Xlex system if you ask them nicely, but you'll have to wait for them to respond to your e-mail.