The concept of dense small cell has been recently emerged as a promising architecture that can significantly improve spectrum efficiency and system capacity.However, it brings frequent handover for user equipment(UE).Furthermore, this will bring a great deal of signaling overhead to the core network.Virtual technology has been received widespread attention for solving this problem.Its essence is to form virtual cells by clustering various terminals properly.The local mobility management proposed recently is based on the virtual technology.Therefore, the formation process of virtual cells is the basis for the research in local mobility management.So clustering scheme for dense small cell network has been studied in this paper, and a maximum benefit merging algorithm based on undirected weighted graph has been proposed.There are X2 interfaces between the cluster head and each of cluster members within the same cluster.The cluster heads manage the handover among cluster members acting as the local anchors.The proposed clustering scheme is useful for local mobility management.The simulation results show that the proposed clustering algorithm can decrease the signaling overhead more than 70% and 20% compared with the non-clustering algorithm and other clustering algorithms respectively. The concept of dense small cell has been considered as a promising architecture that can significantly improve spectrum efficiency and system capacity. However, it brings frequent handover for user equipment (UE) .Furthermore, this will bring a great deal of signaling overhead to the core network.Virtual technology has been received widespread attention for solving this problem.Its essence is to form virtual cells by clustering various terminals properly.The local mobility management proposed recently is based on the virtual technology.Therefore, the formation process of virtual cells is the basis for the research in local mobility management. S clustering scheme for dense small cell network has been studied in this paper, and a maximum benefit merging algorithm based on undirected weighted graph has been proposed. There are X2 interfaces between the cluster head and each of cluster members within the same cluster. The cluster heads manage the switching among cluster members acting as the local anchors.The proposed clustering scheme is useful for local mobility management. The simulation results show that the proposed clustering algorithm can decrease the signaling overhead more than 70% and 20% compared with the non-clustering algorithm and other clustering algorithms respectively.