Environmental variables play important roles in species composition and richness of alpine vegetation. In this study, the characteristics and environmental variables of the alpine vegetation situated at an elevation of about 3350―3750 m on the south slope and an elevation of about 3350―3650 m on north slope of Taibai Mountain of Qinling Mountain ranges in China were investigated in August and September 2009, and August 2010. On each slope, we constructed one transect along elevational gradient. In the transect, 15 typical plots of 2 m × 2 m (for shrubs) at approximately 100 m (fluctuating up and/or down by 20 m at a given elevation) elevation intervals were determined. In the corner of each plot, a 1 m × 1 m subplot was choosed for herbs. Thus, 5 and 4 elevational gradients were set up on the south and north slope, respectively, and totally 135 plots of 2 m × 2 m and subplots of 1 m × 1 m were surveyed. We recorded the community characteristics (including species name, cover, number and height) and the environmental variables (containing altitude, aspect, slope, stone cover, soil depth, soil pH, soil organic matter content, total nitrogen content, alkali-hydrolyzable nitrogen content, total phosphorus content, available phosphorus content, total potassium content and available potassium content). Canonical correspondence analysis (CCA) was used to elucidate the relationship between species composition and environmental variables, partial CCA to calculate the gross and net effects on species composition for each environmental variable, and generalized linear model (GLM) to fit the response of species richness to each environmental variable. Thirteen environmental variables totally explained 31.7% of total variation in species composition, of which, the net effects of altitude, slope, soil alkali-hydrolyzable nitrogen content, total phosphorus content, aspect, stone cover, pH value, soil depth, organic matter content, available phosphorus content and total nitrogen content on species composition were significant (P < 0.05), but their effect intensities decreased successively. Some species, such as Rhododendron deterile, Potentilla glabra, Trollius farreri and Rhodiola kirilowii, were associated with steep slope, hot-dry aspect, organic matter and alkali-hydrolyzable nitrogen rich-soil; otherwise, Primula knuthiana, Saussurea taipaiensis, and Saxifraga gemmigera were more dominant in cold-moist aspect, low acidic and total phosphorus rich- but total potassium poor-soil; Besides, Aconitum tanguticum, Kobresia vidua, Papaver nudicaule, etc. occurred in a high altitude, gentle slope, thicker and available potassium rich-soil. However, Salix cupularis, Carex capilliformis var. major, Ajania variifolia, Ligusticum sinense var. alpinum, etc. were found in most plots since all environmental variables had little effects on their distribution. There were four significant (P < 0.05) relationships detected by GLM between species richness and environmental variables: (i) a monotonic increase with increasing altitude and soil depth; (ii) a monotonic decrease along aspect and soil total nitrogen content; (iii) an unimodal relationship with aspect, pH value, alkali-hydrolyzable nitrogen content and total phosphorus content and (iv) a reverse hump-shaped relationship with soil organic matter content and total potassium content. Among these significant environmental variables, altitude, soil alkali-hydrolyzable nitrogen content, pH value, organic matter content and aspect explained the most variation in species richness.