In order to explore the modeling of the impact of mesoscale eddies on albacore tuna （Thunnus alalunga） catches in the South Pacific Ocean, and to reveal the correlation between different types of eddies and longline catches of albacore tuna in the South Pacific Ocean, in this study, the EddyNet deep learning model was used to automatically identify and detect mesoscale eddies in the South Pacific Ocean based on eddy feature environmental data, and combined with fishery sounding data to analyze the effects of the relative distances from the fishing stations to the eddy edges and eddy centers on the catch per unit effort （CPUE） of albacore tuna. The results show that anticyclonic eddies in the South Pacific Ocean are usually more numerous than cyclonic eddies and have a larger radius than cyclonic eddies. From beyond the eddy edge, both cyclonic and anticyclonic eddies showed a decreasing trend in CPUE for albacore tuna as the relative distance from the fishing site to the nearest eddy edge increased. Within the eddy disturbance region, the CPUE of the anticyclonic eddy was higher than that of the cyclonic eddy, and gradually decreased as the relative distance from the fishing site to the center of the eddy increased. The study suggests that the mesoscale has a positive regulatory effect on catch yield compared to the background field, and that the anticyclonic and cyclonic eddies have different influence mechanisms on the spatial distribution of albacore tuna. This study further deepens the understanding of the complex relationship between oceanic mesoscale eddies and the spatial distribution of albacore tuna, and provides a scientific basis for the optimization of fishing efficiency and sustainable management of fisheries.