As one of the most important laser satellite,Lageos serves for the geodynamic research dedicatedly,and the improvement of its satellite laser ranging data quality is of great significance to the study in geodynamics. Based on the distribution of Lageos laser retro reflectors,the relationship between the echo data of different floors and the reflective depth is deduced,and the extraction model of Lageos leading edge data is established. After applying the model into SHAO′s pre processing,we compared the results of two methods the leading edge method and the conventional method. It is proved that the leading edge method improve the NP data quality. The inner coincidence results show that the standard deviation of the distance between the normal points generated by the extraction model and the reference line is significantly reduced,from 3.37mm to 0.50mm.The external coincidence results show that the average range bias value is smaller and the normal points generated by the extraction model through participating in the rapid and precise orbit determination,is from 28mm to 21mm. Therefore,the stability and accuracy of the normal point could be improved by the extraction model. At the same time,it provides an appropriate thought for producing high quality Lageos satellite laser ranging data.