Abstract: In command and control systems, the processing of battlefield geographic spatial data needs large computation and long time consuming. For solving this problem, the paper proposes an optimized spatial search method based on GPGPU. After designing one parallel grid index which could be applied to GPU environment, the paper provides the parallel spatial search algorithm used CPU and GPU heterogeneous model. The experiment result shows that, the optimized method could improve the performance by one order of magnitude compared to CPU method, which offered a new idea to applying GPGPU in military field.