Real-time monitoring of ship trajectories is significant for ensuring transportation safety and optimizing efficiency in waterway transport. With the application of deep learning, object detection and tracking methods have rapidly advanced. Inspired by the “Tracking-by-Detection” framework, this paper proposes a ship target identification and tracking method based on maritime radar. Firstly, the YOLOv8 algorithm is used to detect ship targets in radar images. Then, using the motion information of ships, the OC-SORT method is applied to associate detection results across frames, ultimately forming ship trajectories. Validation shows that the proposed method achieves a map value of 0.93 for object detection and a HOTA value of 75.62 percent for object tracking, with an overall operation speed of 25 FPS. Comparisons with different methods indicate that the proposed method exhibits good superiority and adaptability in the task of maritime radar target tracking and has potential for real-time applications.