Due to the importance of recommendation system in the e-learning domain and its ability to recommends students with resources of their individual interest, researchers have conducted into RS field and proposed various techniques and approaches to develop and enhance the quality of recommendation systems. In this comprehensive current State-of-the-Art we investigate techniques, similarity measurements, and evolution metrics in the journey of RS revolution. Besides, we have addressed challenges facing RS development such as, cold-star, data sparsity, and scalability. In addition, a wide range of RS applications have been addressed, to provide students with personalized content, increasing student engagements, and enhance learning experience. Also, we addressed trends and new directions of RS that promise to improve recommendations quality and in rich RS improvement such as, feature engineering and feature crossing, context aware, and emotion aware to offer more accurate, responsive, and relevant learning resources.