<p>This paper introduces a polishing force control strategy adopting multi-sensor measurement system to enhance the insufficient of the polishing force control system using a single sensor. The multi-sensor fusion measurement experimental platform driven by compressed air is developed. The measurement process includes the air pressure sensor located at the cylinder inlet, the force sensor located at the piston, and the multi-sensors group arranged below the fixture for polishing workpiece. The measured data of the multi-sensors are fused based on the improved adaptive weighted multi-sensor fusion method based on Kalman filter (short for improved adaptive weighted fusion with Kalman filter) to promote the accuracy of system measurement. Simulations and experiments under different polishing forces are conducted. The results shows that the improved adaptive weighted fusion with Kalman filter can provide more accurate measurements for the control system. Combined with PID algorithm, control error of the polishing force is less than ± 0.1 N. This algorithm can ensure the accuracy of measurement and stability of control under constant force polishing and variable force polishing thereby ensuring the force control precision in polishing process and promoting automation control of polishing. This method is distinctively significant with reference value for CNC machine polishing, robot polishing and other industrial fields.</p>