Although low-energy adaptive clustering hierarchy (LEACH) protocol adopts distributed clustering algorithm and randomized rotation of Cluster Heads (CHs) mechanism to reduce energy consumption, election of CHs without residual energy and position information of each nodes brings about irregular distribution of CH, low network coverage and short lifecycle. In order to avoid these shortcomings, a Grid-based Cluster Head Selection (GCHS) is proposed. Referring to sensing distance, network is divided into several grids equivalent to independent clusters which can meet network coverage and connectivity. Furthermore, CH is selected ground on residual energy level of each node. Simulation taking full advantage of concept mentioned above manifests that it achieves a significant improvement in network coverage and lifecycle.