When considering long-term education, e-Learning is truly innovative when compared to traditional in-person teaching and learning methods. An increasing number of people these days are benefiting from different e-Learning initiatives. The traditional “one-size-fits-all” learning model, which gives every learner access to the same set of learning resources, is put to the test by the great diversity of online learners. In the real world, the students may have a variety of interests. Even if they share a common interest, they may possess varying degrees of skill, so they cannot all be treated equally. It is important to offer a customized system that can adapt to the interests and ability levels of students. Various recommendation techniques can be employed to attain personalization. The overview of observations proposes several techniques for course recommendation in e-Learning platform, which includes (i) machine learning, (ii) natural language processing/sentiment analysis, (iii) deep learning, (iv) hybrid techniques, and (v) learning styles. Out of 80 initially identified papers between 2012 and 2023, for the final synthesis, 33 articles have the significant relation with the course recommendation. The findings raise upon the lack of thorough discussions regarding the importance of recommendation systems for e-Learning in general and online educational platforms in particular, given the unique requirements of e-Learning; the inadequate size of databases employed in specific research; the relevance of recognizing the advantages and disadvantages of every kind of recommender system in a learning environment; and the necessity of more research on implicit feedback compared to explicit learner feedback in order to provide more accurate recommendations.