Abstract: Addressing the challenges of achieving compliant tracking of surgical intentions and ensuring safety constraints for semi-active knee arthroplasty surgical robots, a Model Predictive Compliance Control strategy is proposed. Leveraging the compliant control principles of the admittance model, the design incorporates the explicit constraint-handling capabilities of the Model Predictive Control algorithm to formulate the MPC-based compliance control for the robotic arm. The use of MPC's rolling optimization and feedback correction features enhances control robustness. A comparative experiment with a three-loop PID control is designed to validate the superior trajectory tracking performance of the MPC-based approach, demonstrating its capability to achieve desired compliance dynamics and enhance overall smoothness. Furthermore, the active constraint capabilities of the MPC-based compliance controller are verified on a mechanically complex robotic arm with strong coupling and intricate system parameter structures.