This paper proposes a damping control method for Wave Energy Converters (WECs) based on the Deep Deterministic Policy Gradient (DDPG) algorithm, aiming to maximize energy capture. By adjusting the equivalent resistance of the external load on a permanent magnet DC generator, continuous regulation of the equivalent damping of the Power-Take-Off (PTO) system is achieved, forming a semi-active control strategy that requires no energy feedback. A "wave-to-wire" simulation environment encompassing hydrodynamics, transmission systems, and generator models was established. An analytical solution for the optimal external resistance under regular wave conditions was derived to serve as a benchmark for control performance. Simulation results indicate that, compared to the optimal passive damping control based on frequency-domain analysis, the DDPG controller significantly enhances the system's energy capture efficiency while satisfying constraints on buoy displacement and generator current—achieving approximately a 100% increase in electrical energy output over a 50-second simulation period. This confirms the potential of deep reinforcement learning for achieving end-to-end optimal control in complex wave energy systems.