In order to evaluate the characteristics of complex network in mobile social networks, a modeling method and analytical mechanism on behavior of mobile social network were proposed based on simple messaging service / multimedia messaging service (SMS/MMS) by combining the characteristics of real short message network. At first, social relationship graph was constructed using SMS/MMS. In addition, an analysis mechanism on behavior of mobile social network was presented based on the social relationship graph. The related factors of behavior analysis and their computing model were provided, and the method of social network analysis was used to count and analyze the characteristics of mobile social network, such as out-degree, in-degree, the number of friends, activity degree, intimacy degree, and clustering coefficient. Simulation results show that the real short message network has the characteristics of complex networks and that activity degree can be used to characterize the behavior of users in mobile social networks. The method can be used to evaluate the degree of nodes’ influence in mobile social network.