Abstract: Biomass is an important information in the study of forestry and ecological applications, and remote sensing technology of aboveground biomass estimation in forest ecosystems has attracted intensive attention of the international scholars. Reviewing and discussing different data sources and estimation methods can provide guidance for estimation of forest aboveground biomass. This study discussed the application of single sensor remote sensing data, including optical remote sensing, synthetic aperture radar and LiDAR data in forest biomass estimation, and the advantages of using multi-sources remote sensing data to estimate forest biomass. Then we discussed the traditional analysis methods and machine learning methods (decision tree regression, k-nearest neighbor, artificial neural network, support vector regression, maximum entropy) used for estimating forest biomass. Multi-source remote sensing data integration can combine the advantages of different data and provide rich characteristic information for forest aboveground biomass estimation. Combining machine learning methods is a development trend to improve the accuracy of forest aboveground biomass estimation.