Although conclusions vary widely, there are two suggestions that are applicable to almost any conclusion. Conclusions usually sum up the paper and answer the question "so what?"
When you sum up a paper, you look back over it and ask, "What are the main points I want to reinforce in the reader's mind?" Usually, these will be the same points you promised to cover when you previewed the paper in the introduction, so one of the big challenges is finding a way to hit those points again in a fresh way, hopefully showing that these points have more significance than was apparent in the introduction.
When you answer the question "so what?" you are forced to consider whether you have said anything worth saying. If you have, there ought to be something that "follows from" the point you have made. In other words, the "so what" question forces you to consider the single main point you have made and to explain its significance.
It is not unusual to discover what you wanted to say as you attempt to write the conclusion; that is, often the writing process helps you explore the subject and your views on it, and so when you begin to answer the so what question, you come up with is a conclusion in the sense of "here's what I induce from what I've just said." If you come up with that kind of conclusion, your writing process has been useful for you because it has shown you what your thesis is. In most cases, this kind of concluding statement needs to be moved to your introduction because it is a better controlling statement than the one you started with. After doing that, you should go back through the paper to be sure that the main points you make really support the new thesis, and then you should return to the question "so what?" Don't be surprised if you discover that writing the conclusion sends you back to refine your whole argument.