Abstract: The maritime target monitoring system acquires a large number of trajectory data, deep mining and analysis of the target behavior characteristics is an important way to improve the performance of the traditional maritime target detection technology based on acoustic, optical and electromagnetic characteristics. In the pre-processing of data mining, such as trajectory clustering, trajectory prediction and behavior analysis, the mainstream trajectory segmentation algorithm aims to get the minimum loss and does not give the semantic meaning to the segmented trajectory, which limits the application of segmented trajectory. This paper proposed a segmentation algorithm of maritime target trajectory based behavior pattern. Seven target behavior patterns were defined to achieve the goal of trajectory segmentation. Experiment shows that the algorithm in this paper can segment the trajectory accurately and restore the trajectory accurate, different algorithm parameters correspond to different application scene. Finally, this paper has an outlook on application of segmentation results.