Wireless sensor network combines sensing, computation and communication. Due to limited energy,energy efficiency of sensors is a main concern and a most challenging task for the design of wireless sensor networks. This paper proposes a novel algorithm for network topology, namely Dynamic Energy-Efficient Hierarchical clustering algorithm (DEEH). Different from others, DEEH need to know any local information of sensors. The algorithm can be applied to real large-scale sensor networks in which the sensors have different energy levels and different transmission radius. Compared with the classical clustering algorithm LEACH (Low-Energy Adaptive Clustering Hierarchy), the algorithm is better when the nodes are densely distributed. This paper also considers the selfishness of nodes and analyzes its impact, and introduces a trustful mechanism design that is applied to the algorithm. Under this mechanism, the dominant strategy of selfish nodes is to report their energy truthfully. This strategy can prolong the network lifetime and improve the stability of the network topology.