Optical gyros and star sensors are primary measurement hardware in an attitude control system with high accuracy.The drifts of the optical gyros, however, make an unfavorable impact on the accuracy of the attitude control system.In order for compensations to be provided, this paper presents a least-square method to identify the optical gyro drifts by using flight attitude data from the star sensors and the optical gyros.Equations for identification are formulated by quaternion.Integration of the identification equations and the data from the star sensors are utilized to form a least-square index, in which lower sampling frequency of the star sensors than that of the optical gyros is dealt with effectively.An iterative identification algorithm is presented to minimize the index.Identification procedure under three-axis attitude coupling is illustrated .Simulation results show the effectiveness of the method presented.Proper sample size and sampling frequency are also recommended.