The distributed target estimation problem in a wireless sensor network (WSN) which is under network attack is studied in this paper.Due to the limited measurement range,only some sensors in WSN can measure the target,and at the same time,the nodes are randomly attacked so that the measurement value is injected into false information.An improved consensus Kalman filter algorithm based on the attack detection and recognition strategy is proposed.In this algorithm,firstly,the node judges whether it is attacked based on the attack recognition threshold given in this paper.Secondly,a consensus Kalman filtering algorithm is designed based on the minimum trace fusion principle.Finally,the convergence of the algorithm is analyzed,and a sufficient condition of the attack probability for the boundedness of the mean-square estimation error in WSN is given.Besides,numerical simulations are given to verify the effectiveness and superiority of the algorithm.