This thesis defends the controversial view that results taken from psychology should be used by epistemologists developing their accounts, methodological naturalism, and describes how it should be implemented. I argue methodological naturalism should be implemented by epistemologists adopting an approach to epistemology where they study the goals that can be achieved through epistemic classifications, accepting an account of knowledge, rationality or justified belief only if it describes epistemic concepts in such a way that they can be used by humans to achieve important goals. I show how such an approach complements methodological naturalism by allowing results from psychology to come into the picture to reveal important goals to be achieved. My project begins with a general defence of methodological naturalism. I show how results from psychology provide important insights relevant to epistemology that are not available in the armchair. I respond to a number of common criticisms of methodological naturalism, showing that they are based on misconceptions about the position. I evaluate a number of accounts of the methodology to use in epistemology consistent with methodological naturalism. I evaluate accounts that suggest that epistemologists should retain armchair methods such as conceptual analysis, simply adding a naturalistic component. I reject these on the basis that they cannot avoid some fundamental problems with the armchair methodology. I evaluate the claim that knowledge is a natural kind that can be revealed by consulting intuitions about obvious cases, providing a number of examples showing how implausible the claim is. I argue that knowledge is not a natural kind and instead it is a category required to answer to human needs. Finally, I show how epistemologists can give sufficient recognition to how epistemic concepts are answerable to human needs, avoiding the various criticisms of armchair methods, while adopting a methodology fully consistent with methodological naturalism.