A method is presented for combining GPS and GLAS data to produce a high-quality DEM. First, the GLAS data was transformed so that it could be combined with GPS data to produce a DEM. An algorithm was selected to locate GLAS points and the GPS points covered by GLAS footprints were highlighted. Comparison of the two types of data indicates that the maximum and minimum mean height differences are 1.118 m and 0.997 m, respectively, with a stable standard of about 5-6 cm. The maximum and minimum values became 0.405 m and 0.284 m after rectification by ellipsoid transformation. Detailed components of the error are also discussed. Next, sparse GPS data were extracted as feature points along the GPS measurement profile using a revised angle limitation method. Finally, a DEM was produced for the Dome-A using processed GPS data and transformed GLAS data. 1199 of the GPS points and 53 of the GLAS footprints were used to validate the DEM, with 5 cm and 12 cm as the mean and absolute maximum, respectively, of residuals. Additionally, three other DEMs were produced using original GPS data, original GPS data and GLAS data, and processed GPS data. The DEM produced with processed GPS data exhibited better accuracy than that produced from original GPS data according to comparison of contours and residuals of check points, confirming the importance of the GPS process.