The full waveform data of the large-footprint Geoscience Laser Altimeter System (GLAS) on the Ice, Cloud, and land Elevation Satellite, together with airborne light detection and ranging (lidar) data, were employed to retrieve the basal-area weighted mean height (Lorey’s height) over sloping terrain in the Qilian mountains region, western China. Over mountainous areas with high relief and complex terrain, a GLAS waveform is characterized by multiple energy peaks, which ground and surface objects may be broadened and mixed, making the extraction of canopy height difficult. This study focuses on forests in a mountainous area to derive mean tree height directly from the GLAS waveform information and Gaussian decomposition results. We derived a relationship between the weighted mean tree height derived from airborne lidar data and the predicted mean tree height within the GLAS footprints; the resulting equation explained 82.8% of variance, with an RMSE of 2.8 m. Based on the analysis of different slope categories, it can be demonstrated that the proportion of energy and characteristics of the Gaussian curves greatly influenced the extraction of mean tree height in mountainous areas.