Abstract: To determine the saturation value of Pinus densata forests and Picea-Abies forests biomass, the forest management inventory data and Landsat 8 OLI remote sensing images of the alpine and sub-alpine coniferous forests in the Northwest Yunnan Province were studied, and the saturation points of aboveground biomass (AGB) were estimated by the fitting curves. The linear stepwise regression was used to construct the models which included the non-stratification model, stratification models based on either slope aspects, stand age groups, or the combination of both slope aspects and stand age groups, then the corresponding inversion map of AGB was drawn .The results showed that: ① The saturation points estimated by using Landsat 8 OLI remote sensing images for AGB of Pinus densata forests and Picea-Abies forests were 149.09 t/hm² and 162.3 t/hm² respectively; ② Compared with the non-stratification model, the stratification models had better estimation results, especially for low AGB values (0～30 t/hm²) and high AGB values (>150 t/hm²). Therefore, the stratification models can reduce the impact from the overestimation and underestimation in AGB remote sensing estimation by considering either slope aspects, stand age groups, or the combination of both stratification for Pinus densata forests and Picea-Abies forests to a certain degree, and provide a reference for forest biomass estimation by remote sensing.