A feature point separation multi threshold DP algorithm (FMDP) is proposed to address the issue of missing ship turning and berthing feature information in the Douglas Pucker (DP) algorithm for fishing vessel trajectory compression. The main idea is to filter the feature points in the original trajectory and then integrate the data, maximizing the preservation of feature information in the original trajectory of the fishing boat. Firstly, filter the dwell points in the original trajectory; Secondly, a DP algorithm with angle threshold was proposed to obtain the feature point set of the trajectory; Once again, in order to ensure that the compressed trajectory can reflect the characteristics of different operational tasks of the fishing vessel, a data compression weight mechanism was designed to handle the detention points of the fishing vessel; Then, based on the set of dwell points and feature points, a compressed trajectory image of the fishing boat was constructed to effectively compress trajectories for different operations of the fishing boat. Finally, the trajectory compression effect was verified by comparing the compression rate with the DP algorithm and the influence of weight mechanism on the compression rate of complex and simple trajectories. The results showed that the compression rate of the algorithm was higher than 95%, and compared to the DP algorithm, while maintaining a high compression rate, The feature points of the fishing boat trajectory are well preserved, and the addition of weights enables complex fishing boat trajectories to retain more feature information during the compression process, proving the correctness and effectiveness of the FMDP fishing boat trajectory compression algorithm.