This is a classical philosophical discussion in epistemology.
Beliefs are just things we hold to be true.
Knowledge is a subset of beliefs that we can justify with evidence and arguments.
There are however some beliefs we can know without evidence and argument, namely things like these are my fingers typing this response not someone else's fingers. The reality of an external world, other minds (People), the reality of the past, and continuity of how the world works being the same in the future as it was in the past.
Those are called properly basic beliefs that also count as knowledge despite proof.
For more research on religious epistemology see
http://plato.stanford.edu/entries/religion-epistemology/
Now another issue is how do we know that the criterion for knowledge is correct (how do we judge the evidence) given there must be a criterion for the criterion.
According to Chisholm, epistemology consists of Socratic inquiry into the questions “What can we know?” and “What are the criteria of knowledge?” He thought that a puzzle faces anyone who attempts to answer these questions. It appears that to answer the first question, one needs a criterion to distinguish between things that are known and things that are not known. That is, one needs an answer to the second question. But to have an answer to the second question, he thought, one needs a list of the things one knows so that one can identify the features that distinguish knowledge from its opposite. That is, one needs an answer to the first question.
This problem is discussed in a book entitled, "The problem of the criterion." It may be out on google books.
Next one has the issue of whether they can choose to believe something. I'm a involuntarist but others think and argue that one can choose for more see:
http://plato.stanford.edu/entries/ethics-belief/
Hope this helps.