Aiming at the tracking-shift of traditional target tracking algorithms based on compressed sensing technology， a vehicle target tracking algorithm based on Kalman filtering improved compressed sensing algorithm was proposed in this paper. Firstly， the observed value was obtained by identifying the region with the highest probability of target existence in this frame based on the traditional compressed sensing target tracking algorithm. Secondly， the Kalman filter was used to predict the tracking trajectory of this frame so as to obtain the predicted value， and the Kalman filter gain coefficient was used to correct the predicted value and the observed value to obtain the final target tracking result. Finally， positive and negative samples were taken around the corrected target area to realize the updating of naive Bayes classifiers， and then the real-time updating of target tracking trajectory was achieved. The feasibility of the proposed method was verified by laboratory tests and field experiments. The average tracking error of the proposed method is reduced by 48% and 89%， respectively， compared with the target tracking algorithm based on compressed sensing technology. The tracking trajectory was closer to the real vehicle trajectory.