To address the issue that the deep-sea mining vehicle is vulnerable to time-varying external disturbances caused by some factors such as dynamic ocean currents when operating autonomously on the seabed， and to further improve the motion control accuracy of deep-sea mining vehicle， an inner and outer loop control method of deep-sea mining vehicle based on hybrid LQR/H∞ and GPC is proposed. Based on the error kinematic model of the deep-sea mining vehicle that considers time-varying external disturbances， the LQR and H∞ control methods are combined to construct the azimuth outer loop controller， which enables the controller to possess both the ability of fast tracking of linear optimal control methods and strong robustness. The speed inner loop controller is constructed by combining the GPC speed controller based on the track’s longitudinally kinematics model and the deep-sea mining vehicle’s dynamic model， so that the controller can ensure real-time solutions while considering the dynamic characteristics of the deep-sea mining vehicle. The simulation results show that compared with the commonly used MPC controller in the lateral controller and the traditional PID controller in the longitudinal controller， the proposed method not only has higher lateral and longitudinal control accuracy， but also shows stronger robustness to time-varying external disturbances.