Mobile learning is becoming increasingly popular. Educational web sites can be used as supporting learning tools for students who wish to supplement their knowledge without restrictions of time and place. The continuously increasing demand for enhanced remote and mobile services, as well as the difficulty in easily incorporating current learning services for mobile users, renders essential the adaptation of educational material for these requirements. The objective of this work is to present and evaluate a methodology for producing content semantics from learning material. The proposed approach results in recommending links, which are relevant to the mobile users' interests, by exploiting the recorded usage of an educational portal and the semantics of the learning content. The implementation results reveal enhanced capabilities in mobile learners' web experience and usability.