Abstract: Aiming at spacecraft equipped with large flexible appendages, the moment of inertia of these appendages constitutes a substantial proportion of the total system, resulting in a significantly accentuated rigid-flexible coupling effect. The pronounced coupling effect can easily excites flexible vibration that severely degrade attitude control performance and may even lead to system instability. In order to address this problem, an adaptive robust model predictive control strategy is designed. First, a disturbance observer is integrated with robust model predictive control to estimate the composite disturbance arising from flexible vibration online. Based on this estimation, a nominal system optimization problem is formulated to solve the optimal control sequence, thereby effectively reducing the impact of large flexible vibration on the attitude system. Second, a simplified dual neural network based on neurodynamic optimization theory is employed to solve the optimization problem efficiently, significantly reducing computational complexity and accelerating the response of the attitude control system. Finally, an adaptive sliding mode auxiliary compensation control law is designed to further reduce the error in observing the flexible vibration and other residual disturbance. Consequently, the actual system state closely tracks the nominal state trajectory. Simulation results show that the designed control strategy can effectively suppress vibration of the large flexible appendages while achieving rapid convergence of the spacecraft attitude.