Multi-angle surface reflectance data from the NASA Jet Propulsion Laboratory Multi-angle Imaging Spectro-Radiometer (MISR) were used to map aboveground biomass density (AGB, Mg ha−1) in the forests of the southwestern United States inter-annually from 2000 to 2015. The approach uses a multi-angle index that has a loge relationship with AGB estimates in the National Biomass and Carbon Dataset 2000 (NBCD 2000). MISR Level 1B2 Terrain radiance data from May 15–June 15 of each year were converted to mapped surface bidirectional reflectance factors (BRFs) and leveraged to adjust the kernel weights of the RossThin-LiSparse-Reciprocal Bidirectional Reflectance Distribution Function (BRDF) model. The kernel weights with the lowest model-fitting RMSE were selected as the least likely to be cloud-contaminated and were used to generate synthetic MISR datasets. An optimal index calculated using BRFs modeled in the solar principal plane was found with respect to NBCD 2000 estimates for 19 sites near Mt. Lindsey, Colorado. These relationships were found in areas with AGB ranging from 20 to 190 Mg ha−1, with the model yielding R2 = 0.91 (RMSE: 15.4 Mg ha−1). With spectral-nadir metrics, the R2 values obtained were 0.07, 0.32, and 0.37 for NIR band BRFs, NDVI, and red band BRFs, respectively. For regional application, a simplified single coefficient model was fitted to the NBCD 2000 data, to account for variations in forest type, soils, and topography. The resulting AGB maps were consistent with estimates from up-scaled 2005 ICESat GLAS data and 2013 NASA Carbon Monitoring System airborne lidar-derived estimates for the Rim Fire area in California; and with the 2005 GLAS-based map across the southwestern United States. Trajectories were stable through time and losses from fire and beetle disturbance matched historical data in published sources. MISR estimates were found to reliably capture ABG compared to radar- and lidar-derived estimates across the southwestern United States (N = 11,019,944), with an RMSE of 37.0 Mg ha−1 and R2 = 0.9 vs GLAS estimates.