Abstract: To address the challenges of unmodeled dynamics and external disturbances in robotic manipulator systems, a sliding mode robust control algorithm based on the computed torque method is proposed. The core design comprises employing the computed torque method for feedforward compensation of the manipulator's nonlinear dynamics to establish a nominal linearized system. Sliding mode control is introduced to ensure dynamic convergence of tracking errors along a predetermined sliding surface, utilizing a saturation function instead of the sign function to effectively suppress inherent high-frequency chattering. An additional robust disturbance compensation term is designed to enhance system robustness and suppression capability against lumped disturbances, which encompass external perturbations and model uncertainties. The sliding mode control gain matrix is optimized via a proportional-derivative control law to improve the transient response speed and stability of the closed-loop system. For algorithm validation, simulations were conducted under three typical operating conditions subject to random disturbances. Computed torque method-based sliding mode control and active disturbance rejection sliding mode control utilizing an extended state observer served as comparative controllers. The evaluation framework incorporated four key tracking performance metrics and total system energy consumption. Results demonstrate that, compared to the two comparative methods, the proposed computed torque method-based sliding mode robust control algorithm significantly enhances control performance. Key performance metrics exhibit an average reduction exceeding 10%, alongside lower total energy consumption. This indicates the proposed method's efficacy in improving trajectory tracking accuracy, robust stability, and energy utilization efficiency for manipulators operating in complex disturbance environments.