Abstract: As the explosive growth of global information，the information overload is one of the most critical problems and is becoming more and more serious. Recommendation systems are one of the powerful ways to solve the problem. The definition of recommendation system is first introduced. A comparison study is conducted based on the four main recommendation algorithms: content-based recommendation, collaborative filtering recommendation, knowledge-based recommendation and hybrid recommendation. The evaluation methods, evaluation metrics and recommendation benchmarked datasets are also presented. At last the difficulties and future directions of recommendation systems are given.