The initial value of the two-step method will seriously affect the accuracy of star sensors optical error calibration algorithm.To solve the probleman optimization method of the two-step method based on Artificial Fish Swarm Algorithm (AFSA) is proposed.This method takes advantage of the characteristics of AFSAwhich can easily jump out of local optimum and then conduct global optimization.Before the implementation of the two-step methodthe output of a group of optimal initial values is used for the initialization of the two-step method,and finally the calibration values with higher accuracy are obtained.The simulation results show that the improved two-step method can improve the calibration accuracy and stability by using the advantages of AFSA as well as overcoming the disadvantages of the two-step method.