Plant functional traits reflect the responses of plants to environmental changes and are key elements in studies of the relationship among plant individuals, ecosystem functions and environment factors. The relationship between plant functional traits and environment has become an important topic in phytoecological research. The characteristics of plant functional traits at the community level and their variations induced by environmental changes are important to understand the development of plant communities under different situations and their adaptation strategies. In order to understand how plant functional traits change along the environmental gradients at the community level, and to find out which environmental factors control this process in Yanhe River catchment, we measured nine plant functional traits in typical plant communities along different environmental gradients. Soil water in the field was measured as one environmental factor, and the rest environmental factors relating to climate and topography were calculated and interpolated under the ArcGIS platform. The nine plant functional traits assessed included leaf thickness, specific leaf area, specific root length, leaf density, root density, leaf nitrogen per mass, root nitrogen per mass, seed mass, seed volume. Weighted mean values of the nine traits were then calculated referring to the importance values of different species within communities. Then community-weighted mean trait values were regressed against environmental variables. The results revealed a linear response for the community-weighted mean trait values compared to eight environmental variables. Soil water up to a depth of 200 cm, average annual evaporation and aspect significantly affected the leaf thickness at the community level(P<0.05). The specific leaf area at the community level was influenced mainly by total annual precipitation from July to September. Meanwhile, the total annual precipitation from July to September, aspect and soil water up to a depth of 200 cm significantly influenced the mean leaf density at the community level(P<0.05). Aspect resulted in the greatest variation in specific root length and root nitrogen per mass. Average annual precipitation was the most important factor in determining root density. The average annual temperature from April to October had a significant impact on leaf nitrogen per mass at the community level. Average annual precipitation, total annual precipitation from July to September, seasonal annual precipitation and slope exhibited the greatest influences on the seed mass, however, the seed volume was influenced primarily by the average annual precipitation and slope. Moreover, the nine plant functional traits evaluated at the community level were closely related and strongly correlated to the environment. This results showed that an environment filter was likely involved in the process of community assembly within the study area. For example, in the wet area of Yanhe river catchment, plant communities tended to have low specific leaf area and high leaf density, as well as large and heavy seeds, while plant communities in the dry region exhibited thick leaves, small, light seeds, and produced less aboveground biomass and more underground biomass. These conclusions can serve as an important guide for vegetation rehabilitation in hilly area of Loess Plateau.