Traditional recommender systems provide the user with a list of items supposed to be of interest to the user. Each item is a single independent object and the entire result represents alternatives that match user’s preferences. Another category of recommender systems provides recommendations as collections of items. For this type of systems the recommended items are not alternatives but items to be taken in a “certain order”. In this work, we propose a system that recommends learning courses sequences. Another objective of our system is to enable efficient courses search by proposing an approach based on a multi-criteria weighting method. Our general purpose is to recommend sequences of learning courses using the user’s profile and also to search specific courses by keywords and to suggest related courses. Our goal is to facilitate the learning process and satisfy the user’s need.