Based on the measured biomass data from 44 Larix gmelinii principis-rupprechtii trees in Sumushan Forest Farm of Inner Mongolia,different tree measurement factors were used to construct a single-tree independent basic model and a compatibility model under three methods(total direct control adjustment method,total direct control simultaneous equations method and algebraic sum control method).1)The determination coefficient of the optimal independent basic model and the compatibility model with DBH as the independent variable ranged from 0.766 to 0.943,the root mean square error(RMSE)was less than 20.030 kg,the total relative error was within±1.905%,and the absolute value of the average relative error was less than 16.905%.The prediction accuracy of each biomass model ranged from 93.7% to 95.6%.2)The determination coefficient of the optimal independent basic model,the compatibility model with DBH,and tree height as independent variables was from 0.767 to 0.985,the RMSE was less than 11.636 kg,the total relative error was less than 1.893%,and the absolute value of the average relative error was less than 16.823%.The prediction accuracy of each biomass model was 93.7%~97.9%.3)All models showed a good fitting effect,no significant systematic deviation,strong degree of fit with the sample points,and high prediction accuracy.4)All models could well predict the biomass of Larix gmelinii,var.principis-rupprechtii and the binary model was better than the unary model.The compatibility model could effectively solve the problem of incompatibility between the total aboveground biomass and the components of each organ,and the algebraic and control methods were better than the other two methods.Considering the modeling efficiency and accuracy,it is recommended to use the binary compatible biomass model of algebra and control method as the biomass prediction model of larch plantation in this area.