Compared with learning in classrooms, classical e-learning systems are less adaptive and once a system that supports a particular strategy has been designed and implemented, it is less likely to change according to student's interactions and preferences. Remote educational systems should be developed to ensure as much as necessary the personalization of learning tasks. New e-learning environments should appear to act as intelligent systems that better fit the needs of their users and especially students according to their interests, preferences, motivations, objectives and knowledge. In this paper, we present how the personalization of students' learning process leveraging the use of social semantic web, using resource description framework models, ontologies, social networking and collaborative tagging. Our aim is to develop an approach of personalization according to students' preferences, interests and knowledge by defining for them the best learning paths, this means, provide them as recommendations the best collaborators and the relevant resources that better fit their needs. We present a new method of recommendation based on users' similarity calculation. We demonstrate the effectiveness of our approach through the design, implementation, analysis and evaluation of a social learning environment.