The star and satellite target detection and recognition are one of the important applications of space surveillance system. Because of the characteristic of point target imaging of the map image, and a large number of the interference of background stars, the feature extraction of the map for target recognition is difficult, so the location of the object is the key characteristics to realize target identification. Gaussian curved surface fitting method is one of the target centroid extraction algorithms to be used widely. Theoretical analysis and experiment show that the traditional Gaussian curved surface fitting method of the satellite motion poisoning has much error. So, the anisotropic Gaussian surface fitting model is put forward, and the model by using two different Gaussian blur parameters and rotation factors to capture target of anisotropic characteristics of different directions, which is suitable for the fuzzy random direction caused by the satellite movement. Simulation experiments and real data test show that the overall positioning accuracy of this method can achieve 0.008 and 0.04, respectively, which is able to accurately extract the map target centroid, and improved greatly than that of the traditional methods.