Considering the millimeter wave radar existing more measurement errors because of noise interference on the battlefield, in order to make the data of millimeter wave radar more reliable in the tank in anticollision system, and use the Kalman filter algorithm to track the near goal in front of tanks. The algorithm uses different dynamic model of Kalman filter for parallel processing, thus realize the parallel filtering and update probability of the model on the basis of matching likelihood function,and combines with all filter to obtain state estimation. The algorithm can effectively track the chariot in front of the tank, acquire of the distance, speed and other information. It has the characteristics of high precision, and reduce the false alarm rate.