Abstract: The trajectory tracking accuracy of omnidirectional mobile robots (OMR) is easily affected by system uncertainties and external disturbances in complex environments. To improve the trajectory tracking performance of the MY-3 omnidirectional mobile robot, we propose a sliding mode control (SMC) method based on reduced-order extended state observer (ROESO). Firstly, we establish a mathematical model of the MY-3 robot, and unify the system uncertainties, nonlinear terms, and external disturbances as “total disturbance”, which is real-time estimated and compensated by the ROESO to reduce its impact on the system, while avoiding the inversion of the inertia matrix. Based on this, we design a sliding mode control law, and combine it with the ROESO to enhance the system's robustness and suppress the chattering problem in sliding mode control. Finally, simulations and physical experiments validate the robustness of the proposed method under load variations. The results show that the proposed method has good robustness under load variations, significantly reduces the trajectory tracking error, improves the system’s tracking accuracy, and enhances its adaptability to external disturbances. Meanwhile, the control signal becomes smoother, verifying the effectiveness of the ROESO in disturbance compensation.