Abstract: This paper introduces the basic principle of robot pose matching and two optimization algorithms,i.e.nonlinear optimization procedure and iterative linearization of the equations,and uses the two optimization algorithms respectively to calibrate the geometrical parameters of a polishing robot.Based on analyzing the factors affecting the robot position error,the algorithms eliminate the parameters which are non-sensitive to position accuracy and whose effects are indistinguishable to others,and repeat the calibration process.The simulation results show that calibration effect is not quite different from the algorithms obtained with the complete parameter set.In some cases,the latter is even better than the former.So it is feasible to calibrate the robot by substituting the complete parameter set with reduced set,with an improved accuracy.