Abstract: Based on the detrended correspendence analysis (DCA), and environmental interpretation mathematical models and 136 subalpine dark coniferous forest (spruce and fir are dominant species community samples collested from Northwest Sichuan an South Gansu, the main vegetation types and ecological gradients, and their quantitative relations with environmental factors of the region are given. The data from 40 meteorological stations in this region are used to get the multivarite regression for estimating the climatic information of various Spruce and Fir forests according to the longitude, latitude and altitude of each sample plot- It is shown by the analysis that the vegetation types of spruce and Fir forest and their distribution are mainly determined by the thermal and moisture (including soil fertility) gradients. In moisture gradients, Moss-Abies, Moss-Picea, Rhododendron-Abies and Rhododendron-Picea forests are dominating moist habitat, Sinarundinar-Abies, Sinarundinar-Picea forest in mesotrophic habitat Quercus-Abies and Quercus-Picea are predominant in drier habitat. In thermal gradients, from warm to cold, the vegetation types are Shrub-Picea forest Betula forest--Deciduous broad-leaved forest—Quercus-Picea forest—Sinarundinar-Picea forest—Sinarundinar-Betula forest—Quercus-Abies forest—Moss-Betula forest—Moss-Picea forest—Sinarundinar-Abies forest—Moss-Picea-Abies forest--Moss-Abies forest—Calamagrostis-Picea forest—Rhododendron-Abies forest. The thermal demand of Picea is higher than that of Abies. The moisture demand of Picea-Abies forest in Northwest Sichuan is higher than that in South Gansu.The analysis of spatial distribution of plant communities, key environmental factors and environmental interpretation by mathematical models and are presented in this paper.