The footprints of a satellite laser altimeter have an elevation accuracy of the decimeter order, which satisfies the elevation accuracy needs of ground control points (GCP) for mapping. However, the horizontal accuracy of footprints is only few tens of meters, and only the footprints illuminating on flat ground targets can be used as GCPs. In this paper, the waveform model was derived and used to develop a waveform simulator of laser altimeters. Compared with the current method of waveform matching, the new simulator considered more detailed effects arising from the device and target, e.g., time and spatial distribution of lasers, surface profile, and surface reflectivity. The airborne LiDAR point cloud and the Geoscience Laser Altimeter System(GLAS) data were involved to match the best-fit waveform by maximizing the correlation coefficients of the simulated waveform and GLAS waveform, and the precise location of every GLAS footprint could be acquired where the correlation coefficient was the maximum. The results show that, the mean of maximum correlation coefficients is more than 0.9 during the GLAS operating periods with normal received energies, and the horizontal accuracy of footprints is approximately 2 m after the waveform matching. The proposed method can be used to extract the laser GCPs on complex relief of the surface.