Abstract: The multi-sensor data fusion plays a valuable role in the underwater acoustic signal processing. In this paper, a multi-sensor distributed detection data fusion system based on Neyman-Pearson criterion is presented. The global optimal multi-sensor data fusion system and the local optimal decision rules are studied, respectively. On the condition that all sensors conduct the detection independently, simulations are carried out for triple-sensor data fusion, and the results show that the novel system has better performance.