This paper proposes a distributed acoustic passive localization method using WSN. First, head nodes compute the azimuth angle and pitch angle based on the time delay of arrival (TDOA) from each participating WSN nodes in clusters. Second, the location of acoustic source is calculated from the angles measured above and geometry positions of the two clusters. Then the accuracy of localization estimation is discussed in theory, and the influence factors and localization error are analyzed by simulation. The results show the algorithm can provide adequate precision localization estimation with low computational complexity. The simple algorithms at the node level can be combined to improve estimation and system performance.