Abstract. Recent years, with the growing amount of web resources, collaborative or social tagging systemsallow internet users to manage and share online resources. Online sources service such as Flickr allows toshare their photos with family, friends, and the online community. Based on the real situation, it is necessaryto provide high-quality tag suggestions for sharing related resources. For this goal, several tag recommendationmodels have been developed. However, the results of tag recommendation did not satisfy users' need.Grouping is an attractive interaction metaphor for users to create reference collections of tags they are interestedin. In this paper, by considering semantic analysis and users' interests, the personalized tag group recommendationmodel has been designed for improving the quality of tag recommendation. Finally, we demonstratethe feasibility of the proposed personalized tag group recommendation model in the social photosharing by the experiment.