Based on characteristics of dynamic clustering structure and combined with superiority of particle filter of selecting excellent weight

distributed target tracking algorithm in WSNs is studied.This method uses the improved particle filter（PF） and the transfer relationship between cluster and cluster to obtain the dynamic state of the target.According to current local estimation position

the predicting speed and acceleration are combined to achieve the predicting position.The result shows that this method can not only save more energy than centralized target tracking but also improve the tracking precision about 60 %.