Abstract: The problem of pursuing large marine mammals by autonomous underwater vehicles (AUVs) is studied. Because the maneuverability of the target is much better than that of the tracker, the AUV may miss the opportunity of approaching the target if it plans the pursuing path after the convergence of the target motion estimation value. Inspired by the idea of game theory, a target pursuing algorithm based on short-term planning is proposed to solve this problem. In this method, all feasible behaviors of AUV in a short period are firstly constructed, then the situation that the AUV will reach under each behavior is evaluated, and finally the behavior that makes AUV in the most advantageous pursuing situation is chosen as the result of planning. In order to examine this method, two simulation experiments are carried out: the first experiment shows that the pursuing success rate of the proposed method is 5.2% higher than that of the traditional method, and the second experiment shows that the AUV can still effectively pursue the target even under large measurement errors. The simulation results show that this method not only improves the success rate of pursuing but also has good robustness.