In this paper, ALOS/PALSAR FBD L-band level 1.5 data were used to estimate the above-ground biomass of natural forest in Thua Thien Hue province. The radar backscatter at relatively long wavelengths can penetrate through the canopy and leaves in natural forestry, thus it is suitable for estimating biomass. The methodology was using the regression model to analysis the sensitivity of backscatter to above-ground biomass in natural forest. In-situ data were collected at 60 sites in 4 forest types of volume (rich, medium, poor and non-volume) with cutting 30 trees for measuring and sampling. Mean error and Mean square error were calculated for comparison between the estimated AGB and in-situ AGB. The rich forest displayed a large difference whereas there was a small difference in AGB value in estimated result and in-situ data.