Posting on arxiv is generally considered a good idea if you are worried about competition from similar results. On the other hand, you will give other mathematicians the means to adapt your ideas and possibly prove better, and possibly more abstract results. But hey that's par for the course, and you cannot (or should not want to) hold onto your results forever.
As to where should you submit, I don't think MO users can really tell from the info included in your question. Of course, you should look at journals that have published similar stuff, but isn't this where your supervisor should really come in?