The quantitative estimation of forest biomass provides an important reference for the study of global carbon storage cycle and climate change. Taking the biomass of Masson pine in Zhaoqing city as the research object, single band,band ratio, texture feature and terrain factor were extracted from the SPOT-7 image and DEM data, meanwhile, the optical selection of these factors are carried out and the preferred factors were divided into three categories: spectral feature, texture feature, and integration of spectral and texture feature. The estimation models of Masson pine biomass were built and analyzed based on Partial Least Squares Regression (PLSR) method and three different sets of factors. The results showed that the determination coefficient (R~2), the root mean square error (RMSE), average error (AE) and relative error (RE) between estimated and measured values based on spectral feature and PLSR were 0.81, 15.30 t/hm~2, -5.65 t/hm~2, 9.30%, the R~2, RMSE, AE and RE values of model based on texture feature and PLSR were0.84, 15.69 t/hm~2, -4.73 t/hm~2, 7.78%, the R2, RMSE, AE and RE values of model based on integration of spectral feature and texture feature and PLSR were 0.83, 15.24 t/hm~2, -5.27 t/hm~2, 8.67%. Therefore. The Masson Pine biomass model constructed based on texture feature was better than the other two methods, and the higher estimation accuracy was also acquired.