According to the characteristics of recommended items in a group recommendation system, i.e., multi-granularity, hesitantcy, fuzziness, firstly, based on the concept of multi-granular hesitant fuzzy linguistic term sets(MHFLTSs), the concept and calculation formula of the multi granularity hesitant fuzzy linguistic entropy are defined, and the attributes of recommended items can be calculated by using MHFLTSs’ entropy measures. Then, the traditional VIKOR method is extended to the multi-granularity fuzzy domain, and the compromise solution formula is improved. The improved VIKOR method is applied to the group recommendation. Finally, the VIKOR method and TOPSIS method are compared from three aspects of theoretical analysis, numerical calculation and sensitivity analysis. The results show that the proposed method is reasonable and effective in the application of group recommendation.