This paper studies the problem of false external data injection information transmitted through wireless sensor networks. Firstly, this paper models the nonlinear system and improves the unscented Kalman filter to solve the problem of false data injection in the transmission process replaces the abnormal innovation data by setting the innovation confidence interval in advance and deduces the error covariance of this method. Then the stability of the designed algorithm is verified. Finally, the effectiveness of the filtering algorithm designed in this paper is illustrated by tracking the state of the system and the mean square error of the simulation system.