Network scalability and energy efficiency in wireless sensor networks should be supported by elaborate topology control. Clustering algorithms are useful topology management approach to reduce communication overhead and exploit data aggregation in wireless sensor networks. A clustering algorithm based on cell combination is presented in the paper for the networks where sensor nodes are distributed densely and the energy of nodes is always limited. In the algorithm, the monitoring region is divided into hexagon cells by utilizing the geographic location information of sensor nodes. Each cluster is formed by at least seven hexagon cells and it shapes nearly like a circle to improve channel reuse and energy efficiency. Nodes with the same cluster identity form a cluster and the cluster head in each cluster is elected from the central cell of each cluster. The simulation results show that the proposed algorithm provides high energy efficiency and low latency and prolongs network lifetime efficiently.