Abstarct: Underwater robots are advanced robotic systems developed for performing exploratory, industrial, and environmental missions in the deep sea. These robots, equipped with multiple degrees of freedom and the ability to integrate mechanical arms, can perform tasks such as inspection, offshore equipment repair, scientific sampling, and rescue operations with high precision. Among this category, the Underwater Vehicle with Dual Manipulators (UVDM) has gained a special place in marine engineering due to its ability to perform multiple complex actions simultaneously and maintain stable interaction with the environment. However, the highly nonlinear behavior, dynamic coupling between the arms and the body, parameter uncertainties, disturbances caused by ocean currents, and actuator saturation limits make the design of a robust and fast controller a significant challenge. In this research, two new control strategies baxsed on sliding mode control theory and disturbance observer are designed and analyzed to achieve robust finite-time control and precise adaptive performance in the presence of uncertainties and disturbances. In the first strategy, an adaptive sliding mode controller baxsed on a model (ASMC) is presented, where the upper bound of disturbances is modeled as a function of the system's state variables, and its coefficients are estimated online using stable adaptation laws. This structure ensures finite-time stability and a significant reduction in chattering without the need for a separate disturbance observer. In the second strategy, a model-free disturbance observer-baxsed robust controller (MFDOC) is developed, which estimates disturbances in real-time without relying on an accurate dynamic model of the system and considers the effect of actuator saturation within the control structure. This method offers lower computational complexity, greater robustness against model changes, and higher implementability in practical environments. The stability of both control structures is explicitly proven using Lyapunov analysis. Simulation results show that the model-baxsed approach provides accurate tracking and appropriate energy consumption, while the model-free structure offers stable and robust performance in the presence of severe disturbances, dynamic uncertainties, and actuator limitations. The findings of this research contribute effectively to the development of robust, adaptive, and finite-time control systems for advanced underwater robots and can be applied in future industrial, exploratory, and environmental applications.