This project proposes a recommendation model for educational content based on the context of a user, which uses a context model that incorporates the role, tasks, programming exercises and their application to the problem of recommendation. The recommendations are made on the basis of the estimate of the difference between the current level of knowledge of a user in front of the skills required in their work context. The experiments developed in the context of the student, show that, using a model of probabilistic reasoning helps to get better recommendations of educational content, according to the missing competences of a student on an issue that needs to learn, which seeks to standardization for recommendation systems.