To address wind power’s inherent intermittency and ensure grid-connected converters’ rapid provision of virtual inertia/damping support during sudden wind speed changes, this study proposes a DDPG-based adaptive optimization strategy for virtual synchronous generators (VSG). The operational principles of grid-connected energy storage VSG are analyzed, with particular emphasis on virtual inertia/damping’s impacts on active power and frequency dynamics. A DDPG-driven adaptive control framework enables real-time parameter adjustments, optimizing power-frequency response. Simulation results demonstrate 0.17s faster regulation, 31.2% lower overshoot, and 0.165Hz reduced frequency fluctuations compared to fixed-parameter control, confirming the strategy’s effectiveness in suppressing oscillations and enhancing system stability.