In order to reduce the interference of atmospheric turbulence and background light on the beacon light,and reduce the influence of turbulence on the optical axis alignment accuracy,optical axis alignment algorithm is studied based on the analysis of Gaussian beam model.First the background light model was analyzed,background noise was filtered according to image model dynamic,then horizontal and vertical mathematical statistics was made according to the beam energy distribution.The spot region was judged by Bayesian discrimination,the spatial filtering of the laser optical axis was achieved in the large field of view.Finally,following the least-squares principle,spot image fitting curve was gained by the Gaussian fitting.And then according to the dynamic curve of each frame image,the extraction of the Gaussian centroid was completed,the optical axis detection accuracy was further improved.The accuracy of the spot image was confirmed by 6.5 km field experiments,the obtained spot image accuracy is 18 μrad,it is about 12 μrad higher than traditional centroid fitting accuracy under the same conditions.