The advances in the educational field and the high complexity of student modelling has provoked it to be one of the more investigated aspects in Intelligent Tutoring Systems (ITSs). The Student Models (SM) should not only represent the student's knowledge, in a wide sense, but rather they should be, insofar as it is possible, a snapshot of the student's reasoning process. In this article, a new approach to student's modelling is proposed that benefits of the Ontological Engineering advantages, so widely used at the present time, to advance in the pursue of a more granular and complete knowledge representation. The goal is to define an ontological basis for SMs characterized by a high flexibility for its integration in varied ITSs, a good adaptability to the student's features, as well as to favor a rich diagnostic process with nonmonotonic reasoning capacities, allowing the treatment of the contradictions raised during the student's reasoning and diagnosis.