Abstract: An account on the application of multivariate analyses to forest soils, including multiple regression, discriminant analysis, cluster analysis, factor analysis and canonical analysis thar have been used by domestic and foreign researchers, is made here.In China, the use of multivariate analyses to soills began in the late seventies. From the very start, research efforts on quantitative and numerical matters have been made in combination with the classification of forest soils and their moisture systems, the evaluation of the productivity of forest soils, and the diagnosis of forest lands with inhibiting soil factors. Major investigations made so far include: the discrimination between podzolic soil and dark brown forest soil in Northeast China, the numerical classification of drab forest soil in North China by fuzzy cluster method, the grouping of moisture regimes in yellow brown soils under Chinese fir plantations near Nanjing by hierarchical clustering, the discriminant analysis of the inhibiting soil conditions of Chinese fir stands in Jiangxi Province, the evaluation of productivity of the soil under Chinese fir and other tree species in southern China by multiple regression and other methods.