Abstract: In the ship tracking task, the surface dragging interference and target occlusion affect the result of ship tracking. Therefore, a long-term ship tracking method that suppresses stern drag is proposed. We pre-processed the image to improve image clarity. We used the offline training dragline detector to detect the location of the water surface dragline, and corrected the ship tracking frame in real time. The object occlusion was judged according to the perceptual Hash algorithm. We used an online detector to detect the target area, retrieved the ship's position and initialized the tracker to achieve long-term tracking. Experimental results prove that this method can effectively suppress the influence of stern drag on tracking, improve ship tracking accuracy, and achieve long-term ship tracking.