Recommender systems have been developed in variety of fields, including music recommender systems which are one of the most interesting ones. Because of the information overload and its varieties in music data, it is difficult to draw out the relevant music. Therefore, recommender systems play an important role in filtering and customizing the desired information. In the proposed system, association rules mining technique is applied to filter the similar users, and then the related songs are recommended to them. For finding the similar users Frequent Pattern Growth Algorithm is being used. Since, FP-Growth allows frequent itemset discovery without the generation of candidate itemset. After finding frequent patterns then association rules are generated. Finally recommendations are done to the users and the accuracy is checked.