In this paper, we present a framework designed to enable peer-based intelligent tutoring, inspired by McCalla’s ecological approach [4]. In particular, we focus on the question of how to sequence the content that is presented to a student and offer an algorithm that selects appropriate learning objects (webpages, videos, research pages etc.) from a repository, for each student based on the previous experiences of similar students. We argue that in order to design an effective education environment for students, it is possible to model each student as an agent, in order to compare that student with previous students who have encountered the system (so, other agents). This perspective is emphasized in our work in particular because we choose to validate our current model using a simulation of all existing students and their learning experiences. As we have been able to gain some benefit through our simulations, we propose that multiagent systems researchers explore the approach of peerbased learning as part of their educational systems.