A new multi-source detection model was proposed based on Neyman-Pearson criterion to reduce the computa-tional complexity caused in the multi-source localization.The Fisher criterion was employed to divide sensors into two parts

where two sources were present and each part corresponds to one of the sources.The WSNAP（weighted subtract on negative add on positive） multi-source location algorithm was applied to localize the multiple sources.The simulation results show that Fisher criterion is able to divide the alarmed sensor into two parts with relatively higher accuracy.The proposed WSNAP has better estimation accuracy than AP（add positive） algorithm and CE（centroid estimator） algorithm under the circumstance of lower computation complexity.Finally

the results are verified using the database of distributed wireless sensor networks.