Applying Virtual Reality (VR) in combination with Intelligent Tutoring Systems (ITSs) is a promising approach to computer based learning and training. However, the classical structure of ITSs has not been conceived to deal with the new sources of information and interaction provided by VR environments. The resulting structures combining both technologies often renounce to the traditional modular separation of ITSs in pursuit of increasing efficiency, but making very difficult the reusing of their components.

In this paper it is described the extension of the classical ITS structure proposed in the MAEVIF model to build reusable and adaptable VR based ITSs.