Abstract: In order to resolve the maneuvering problem in target tracking, an algorithm based on interacting multiple model(IMM) method was presented. In this method every sub-model has its own model match probability that changes with the target's estimated state and measures. The sub-model can soft-jump between each other under the control of Markovian switching coefficients. From simulation it can be seen that the IMM method can improve the tracking accuracy of maneuvering targets.