Abstract: This paper proposes an energy-adaptive clusterhead selection algorithm,in which additional parameters including the remaining energy level of candidates and their energy cost for transmission are synthetically considered to optimize the process of cluster-head selection.The main idea of our algorithm is to avoid choosing nodes with lower residual energy as cluster-heads,thus to ensure balanced energy-usage among the nodes in the network. Preli-minary simulation results show that our protocol greatly improves the performance and efficiently prolongs the network(lifetime) by about 30%.