In this article, we will research the Recommender System's implementation about how it works and the algorithms used. We will explain the Recommender System's algorithms based on mathematical principles, and find feasible methods for improvements. The algorithms based on probability have its significance in Recommender System, we will describe how they help to increase the accuracy and speed of the algorithms. Both the weakness and the strength of two different mathematical distance used to describe the similarity will be detailed illustrated in this article.