There has been a traditional method of calculating word count (as described here by Pam McCutcheon), and there is of course your word processor's word count.

Now I have heard quite a few people say that you must not use the word processor count but the traditional method of calculating it whenever communicating your word count to a publisher. For those of you currently dealing with editors or publishers, have you found this to be the case? Or is it a matter of someone saying so, and thus it keeps getting repeated ad infinitum? Has the technology become prevalent enough that, in reality, publishers are just working with word processor word counts when evaluating manuscripts, or are do they really all stick with the traditional method?