An error compensation and parameter identification method based on sine function and particle swarm optimization was presented to improve the measurement accuracy of circular grating angle sensors. The measurement errors of the sensors were calibrated discretely by a photoelectric autocollimator and a metal polyhedron. By analyzing the calibration data with the Fast Fourier Transform(FFT), it was found that the measurement errors of the sensors are composed mainly of the sinusoidal signals with different frequencies. Thus an error compensation model consisting of seven constants to be determined was presented based on sine functions.Furthermore,by taking these constants as the location coordinates of particles and the average error as the fitness function, one identification model based on particle swarm optimization was built to calculate the constants in the compensation model.Finally, the compensation method was used to compensate the errors of the sensors in an articulated arm coordinated measuring machine. The experimental results show that the average errors of the sensors are reduced about 398~1 102.5 times after compensation.