The invention provides a socialized video recommendation system and method based on implicit community discovery, and belongs to the technical field of socialized video recommendation. The socializedvideo recommendation system comprises a bullet screen comment collection module, an LDA theme clustering module, a random walk module, an implicit community discovery module, a group recommendation module and a result display module. The method comprises the steps that firstly, bullet screen data of a user are crawled from a video service website to be processed to generate a video theme distribution matrix and a theme word distribution matrix; then, a ternary graph method and a clustering method are adopted to obtain a preference matrix of each implicit community for the theme, the similaritybetween each candidate video and the preference matrix is calculated, and a video recommendation sequence is selected according to the similarity sequence; according to the socialized video recommendation system and method, the implicit community recommendation model is established, the bullet screen text information is combined, the implicit community composed of the users is mined, the data sparsity can be relieved, and the socialized video recommendation accuracy is improved.