This paper presents an algorithmic solution for depth control. Depth control for Underwater Vehicle-Manipulator System (UVMS) is essential to complete a specific underwater manipulation. A large number of algorithms have been designed to tackle the depth control problem. However, because of water currents, algorithm for depth control is required strong robustness. This work describes an adaptive sliding mode control (SMC) method to implement the depth control. The designed SMC method combining the proposed adaptive algorithm can alleviate influence of uncertain parameters of system and external disturbances. The stability is proven by a Lyapunov function. Then, we build a ROS-based underwater simulation environment-UWSim. Real-time depth and velocity of UVMS can be obtained by depth gauge and IMU via ROS nodes. Finally, we conduct a depth control simulation with our proposed hybrid-driven underwater vehicle-manipulator system (HD-UVMS). A comparative simulation is also implemented to verify the effectiveness of the proposed method.