The quality of coordinate transformation is influenced by geometric position distribution and measurement errors of common points. In order to improve the effectiveness of coordinate transformation， a method for selecting common points that takes into account position distribution and measurement precision is designed. Firstly， robust estimation method is adopted to process the data and eliminate gross errors. Secondly， centroid the remaining homonymous points， divide them into multiple sets of points based on different quadrants， calculate the projection length of the homonymous points on the central axis of their respective quadrants， and sort them. Proportionally select the points with larger projection lengths in the point sets as the points to be sampled. Thirdly， traverse the set of points to be sampled and extract homonymous points， calculate the conversion precision and coefficient matrix condition number of each group of points. Finally， take the combination with the highest internal compliance accuracy among the smaller number of conditions as the common points. The performance of the method is verified by simulation experiments and case analysis. The results show that compared with references ［24］ and ［28］， the proposed method has better compliance accuracy in both internal and external. The quality of coordinate transformation of this method is higher， and it has robustness.