In this paper, an efficient grid-based sense scheduling method for event detection in Wireless Sensor Networks is proposed. After grid-partitioning sensors in target environment, the sensors within each grid elect one representative to carry up the sense task in turn, which has low communication overhead, balanced energy consumption and good scalability. It is shown by analysis and simulation that the grid-based sense scheduling can achieve energy saving proportional to the node density and better detection delay than random scheduling. The method also provides means to balance network lifetime and detection quality by adjusting size of grids and duty time of nodes.