Variations in ecosystem function in response to land-use changes may be expected to reflect differences in the functional traits of plants. In this study, we sought to reveal the relationship between trait variability and grazing management on typical steppe in Inner Mongolia, and explore the implications of this relationship for ecosystem functioning. We measured aboveground biomass and 18 functional traits of the most abundant plant species in a grassland subject to three grazing-management regimes: long-term grazing, short-term grazing exclusion (since 2008) and long-term grazing exclusion (since 1956). Principal component analysis of the variation in species-level traits revealed trade-offs between the traits that enabled rapid acquisition of resources by fast-growing annual species and those that promoted conservation of resources by perennial grasses, especially Stipa grandis. However, there was no systematic pattern of intra-specific variation in trait values recorded among sites. Aggregation of plant functional traits to the community level revealed a gradient of responses of typical steppe to grazing exclusion. Long-term grazing favoured species whose traits indicate low forage quality and relatively low growth rate. Exclusion of grazing for several years favoured species whose traits indicate relatively high growth rate and high capacity to acquire resources. Exclusion of grazing for several decades favoured species whose morphological and physiological traits indicated low growth rates and high capacity for resource conservation. These community-level traits imply that ecosystem carbon and nutrient stores will change in response to the grazing regime. Long-term grazing will result in decreased plant carbon and nitrogen content, and will lead to carbon and nutrient loss, whereas short-term and long-term grazing exclusion are beneficial to the recovery of carbon and nutrient storage. The findings support the value of community aggregated traits as indicators of environmental or management change and for explaining changes in ecosystem function.