Aiming at the problem of large tracking error of maneuvering targets under nonlinear observation conditions, an interactive multimodel (IMM) and unscented information filtering (UIF) method is proposed. By using the unscented transform (UT) and information filtering (IF), the linearization error under the nonlinear estimation and the direction evaluation of the Jacobian matrix are avoided, and the state estimation of the target motion is realized by analyzing the probability changes of different models in the model set. The simulation results show that compared with IMMExtended Kalman filter (EKF), the proposed IMMUIF target tracking method improves the estimation accuracy and completes the tracking of the target well.