Abstract: To address the inherent strong nonlinearity, parametric uncertainty, and severe external disturbances of un-derground coal mine manipulator systems, which degrade control accuracy and system stability, this paper proposes a composite control method combining an Extended State Observer (ESO) and Sliding Mode Control (SMC). Firstly, the dynamic equation of the manipulator is established using the Lagrangian method, and the lumped system disturbances are modeled as extended states. Secondly, an ESO is constructed to accurately estimate unknown disturbances and implement real-time feedforward compensation. Based on the observed states, a sliding mode control law is designed, and a saturation function is employed to alleviate sliding mode chattering. The Lyapunov stability criterion is utilized to theoretically verify the effectiveness and asymptotic stability of the proposed control strategy. Comparative simulations are conducted on an underground coal mine manipulator, and the results demonstrate that the presented ESO-SMC scheme outperforms conventional SMC in terms of sliding surface convergence rate, disturbance rejection capability, and chattering suppression. The proposed method can effectively adapt to the harsh and complex underground coal mine environments, thereby improving the robustness and tracking accuracy of the manipulator control system.