The accurate estimation of forest height is very important for understanding forest biomass and forest vertical structures. To investigate the potentials of forest height mapping for future Chinese satellite mission concepts with a waveform Lidar system onboard, a field campaign was designed and implemented in Weihe forest farm, Northeastern China in August of 2016. The method we proposed in this paper is that firstly we generate simulated waveforms from Unmanned Aerial Vehicles (UAV) Lidar data, then we use random forest (RF) to get the most relevant variables from 18 waveform parameters driven from our simulated results and 11 terrain parameters from ASTER-DEM. Finally, we used Cubist machine learning algorithm to establish the relationships between 4 different forest heights and the selected variables. Initial results demonstrated that the simulated waveforms could estimate forest height very well.