Aimed at the problem of detecting and characterizing community structure is one of the outstanding issues in the study of complex network, a new community detect algorithm based on Force-directed model was proposed, which categorizes the point to community decided by the force between them. An algorithm named Force-Directed Community Detect, FDCD, and an implementing algorithm using four different Force-directed Models were designed. The experiments show that the algorithms can find community in real social network with high Q Modularity, and each efficiency of algorithm based on LL model reaches the complexity degree of linear computation, which proves fit for the community detection in large network.