In this study, the canopy height estimation over French Guiana was analyzed using multiple linear regressions and the Random Forest technique (RF). This analysis was based on LiDAR waveform metrics extracted from the GLAS (Geoscience Laser Altimeter System) spaceborne LiDAR and terrain information derived from the SRTM (Shuttle Radar Topography Mission) DEM (Digital Elevation Model). Results showed that the use of statistical models based on GLAS waveforms and DEM metrics provides better canopy height estimates in comparison to that obtained by the direct method (RMSE between 3.7 and 4.9 m against 7.9 m with the direct method). The best statistical model is defined as a linear regression of waveform extent, trailing edge extent, and terrain index. Random Forest regressions showed that the waveform extent was the variable that best explained the canopy height. In addition, the estimation of GLAS canopy height by RF using only the waveform extent showed an RMSE of 4.4 m. The best configuration for canopy height estimation using RF used all the metrics: waveform extent, leading edge, trailing edge, and terrain index (RMSE=3.4 m). In our case of low relief area, the use of one or two metrics among the three used in this study in addition to the waveform extent showed a slightly lower precision on the canopy height estimation (RMSE=3.6 m). In conclusion, multiple linear regressions and RF regressions provided similar precision on the canopy height estimation.