In this paper, we proposed a recommender system using energy distance. The energy distance measure is used to detect the degree of mismatch or incompatibility between items in the recommender system. The implementation method of distance correlation is to calculate the distance between users not based on the rating value of each pair, but it focuses on the distribution of each rating value of the first user with all the rating values ​​of other users. Experiments are conducted on the Jester5k dataset, using the data partition method of “split”, “cross validation”, and methods such as Precision and Recall are also selected to evaluate the performance of the recommender models. The results show that the precision value of the proposed model is higher than the compared models, which means that the proposed model gives more suitable and better suggestions to users than the compared models. Besides, the balancing ability of the energy-populating filter recommender model is also higher than the compared models.