If you are writing a journal article, most likely you will not need to reduce it to cover the page limit. Many will charge for additional pages, but the numbers will be much smaller than what you pay to conferences. Try to convince your supervisor and pay the extra charges. If you really have to reduce the page count, the guideline is "leave enough details so that someone else can read the paper and repeat what you did." You can also reshape figures so that they take less space.
If you are writing a "full" or "long" conference paper, you can leave 10-20 references and remove the rest (but bear in mind my previous guidelines on citations :-p). You can also get rid of or shorten the discussion of results.
For a "short" or "poster" paper, you don't need a separate section on related research. If you have one, merge that to the introduction. If the paper should be two pages or less, you can get rid of the "Future Work" section, too. Some short papers do not have an abstract; a well-written introduction can eliminate the need for one.
Once you reduce the paper, leave it aside for half a day and read it again. You will see quite a lot of mistakes, get rid of them. You may now send it to your supervisor for his feedback. To go off-topic, never send a paper too early. Two to five days before the deadline is the best :o).
So much for now. from now on, there will be shorter posts that will be more useful for your survival in grad school and faster graduation :o).