Abstract: To overcome the shortcomings of conventional proportional-integral-derivative (PID) control in suppressing drillstring stick-slip vibration, such as response delay, large overshoot, and insufficient robustness, a model-predictive-control (MPC)-based controller was developed based on a two-degree-of-freedom lumped-parameter drillstring model and the Karnopp friction model. Through multi-objective segmented weighting-factor design and constraints on control torque and bit rotational speed, rapid response, smooth convergence, and high-precision tracking of bit rotational speed were achieved. The proposed method was compared with conventional PID control and fuzzy adaptive PID control. The results show that, under constant operating conditions, the convergence time of the MPC controller is shortened by more than 30% compared with PID control, and the maximum overshoot is reduced by approximately 16% compared with fuzzy adaptive PID control. Under disturbed conditions involving step changes in the desired rotational speed and weight-on-bit, the speed-tracking overshoot of the MPC controller is reduced by approximately 20%, and the transition process is smoother, indicating stronger robustness and disturbance rejection capability. The proposed MPC control method can effectively suppress drillstring stick-slip vibration and maintain favorable dynamic tracking performance and control stability under complex operating conditions, providing a theoretical basis and technical reference for active drillstring vibration control in deep and ultra-deep wells.