Based on the data of 1400 sample plots from continuous forest inventory in Beijing,the multivariate regression models between forest volume,biomass and factors including dominant species,age group,mean diameter,stem number,and canopy closure were developed using the mixed model approach. The results showed that the determination coefficients (R²) of the multivariate mixed models for forest volume and biomass were more than 0.8,the mean prediction errors (MPEs) were less than 3%,and the mean percent standard errors (MPSEs) were less than 25%;and for the forest volume and biomass estimates of 10 forest types,the MPEs were less than 15%,and the MPSEs were less than 30%. It is totally feasible in practice to develop multivariate regression models between forest volume,biomass and quantitative,qualitative/indicative factors of forest stands;the data of sample plots in arboreal forest from 9^th national forest inventory can be used to develop multivariate mixed forest volume and biomass models,which would provide important basis for realizing the decomposition of total provincial data to city,county and each sub-compartment on forest map.