In offshore wind power grid-connected systems, onshore converter stations based on modular multilevel converters are prone to disturbances under operating conditions such as sudden variations in wind power output, equipment switching, and grid voltage sags. To address this issue, an optimized control strategy integrating slope passivity-based sliding mode control and model predictive control (SPSM-MPC) is proposed. The strategy is centered on inner current loop control. A passivity-based sliding mode (PSM) controller is first employed to establish the basic control framework, upon which a slope regulation mechanism is introduced to construct the slope passivity-based sliding mode (SPSM) control strategy. Meanwhile, model predictive control (MPC) is organically embedded into the modulation algorithm framework. Simulation models are used to compare the dynamic performance of three control strategies: conventional PI, PSM, and SPSM-MPC, under normal system operation and three typical disturbance conditions. The results show that the SPSM-MPC strategy reduces the steady-state output current THD to 4.45%, shortens the response time to sudden wind power variations to 2 ms, and reduces the active power stabilization time to 0.15 s under grid voltage sag conditions. Through the synergistic effect of the slope mechanism and predictive control, the SPSM-MPC strategy effectively enhances system robustness under dynamic disturbances, providing a new control scheme for the stable operation of offshore wind power grid-connected systems.