Abstract: Traditional control algorithms do not adequately account for the distinctive maneuvering characteristics of unmanned sailboats,resulting in low path-tracking accuracy and slow convergence under complex operating conditions such as close-hauled (upwind) and beam-reach (crosswind) sailing. To address this issue,this paper proposes a data-driven strategy that integrates intelligent decision-making with robust control and develops a hybrid Proximal Policy Optimization-Proportional-Integral-Derivative (PPO-PID) control framework incorporating a Return-Based Prioritized Experience Pool (RB-PEP). In this framework,an improved PPO algorithm is employed to design a guidance law that outputs discrete heading commands,while a PID controller adjusts the rudder angle in real time to track the commanded heading.Meanwhile,an adaptive reward function is constructed based on heading error,cross-track error,and relative wind direction to enhance the sailboat' s adaptability to different wind conditions. Simulation validation is conducted through random-path training under multiple wind directions and multi-segment path-tracking tasks. The results demonstrate that the proposed PPO-PID algorithm outperforms the standard PPO-PID and classical Line-Of-Sight (LOS)-based control methods in multi-segment path tracking,achieving higher tracking accuracy and smaller errors. The proposed method provides a feasible solution for autonomous navigation of unmanned sailboats in complex sea conditions.