The invention belongs to the technical field of multimedia and in particular relates to a method and a system for recommending music. The method comprises the following steps of: a, accessing a page server through a mobile terminal, browsing page information, and requesting a resource server to acquire page data which is required, wherein the page data comprises music resources and music recommending data; b, recording behaviors of users on accessing of the page server, and recording music acquisition behaviors of the users on the resource server and music resource information which is acquired; c, acquiring behavior records of the users, and classifying the users according to the behavior records; and d, classifying songs in a song classification mode, and generating corresponding song recommending lists according to different user types. By adoption of the invention, the users can more quickly and more effectively listen to favorite songs, so that personalization requirements of the users are met, and a contradiction between mass of the songs and the personalization of user favor is solved.