Your question is too broad: the entire field of computational linguistics is about how to process natural languages on a computer. There are many goals, and for each there are many techniques and possible projects.

For a Master's thesis, you will probably not be doing completely new and groundbreaking work, because you won't have time. It will be better to find some existing research (=read a lot!) and try to extend some of that or improve the techniques.

There are many different projects you could try. One popular example is something about Machine Translation. But each of these areas represents an entire field of research with decades of history to look into.

Good luck.