Excessive farming has intensified wind erosion desertification; this has become a very important factor having serious impact on local economic development and ecological conditions in northern China. This also poses a serious threat to the ecological status of the inland areas of China. In the view of their low tolerance for environmental change, complex community composition, and sensitivity to artificial disturbance, dung beetles can be used to reflect or monitor environmental changes. For anthropogenic or natural environmental changes, plant and mammal community composition are factors that are easily changeable. The abundance and species composition of dung beetle communities in an area can quickly reflect changes to the environment. For the farming-pastoral ecotone, restoring cropland to grassland is an effective way to prevent wind erosion desertification and to recover grassland functions to check wind and prevent sand drifting. From May to September 2006, different methods of restoring grassland from cropland were used in sample plots, such as planting Medicago sativa, planting M. sativa+Artemisia sp., planting Helianthus tuberosus, planting H. tuberosus+Artemisia sp., abandoning cropland, and using existing cropland as a control; the effects of these grassland restoration treatments on dung beetle assemblages were analyzed. The results of such studies potentially could provide the basis for evaluating the ecological effects of such grassland restoration, assist in the perfecting of grassland restoration techniques, and provide information related to the conservation of dung beetle biodiversity in the study region. The pitfall trap method was used for collecting dung beetles. Plastic buckets (25 cm in diameter at the top and 1000 mL in volume) were used for traps and buried in the soil with the top of the trap at ground level. The traps contained 250 mL water, an adequate amount of liquid detergent and insecticide. A nylon bag with 250 g of fresh horse dung was suspended over the top of the trap by iron wire. We set up five traps, spaced 2 m apart in the geographic center of each plot. In each sampling period the traps were set up between 8:00-9:00 a.m. After 5 hours, trapped dung beetles were collected. We randomly captured a total of 21,671 dung beetles, belonging to three families, five genera and 25 species. The dominant species were Aphodius rectus and Aphodius sp.7, which made up 59.34% and 11.71% of the total collected beetles, respectively. All tested grassland restoration measures increased species richness, biomass and the abundance of the dung beetle assemblage, in a seasonally dependent way. We observed a positive correlation between the length of rehabilitation period, average grass height and plant community cover of plots with the species richness, biomass, and abundance of dung beetle assemblages, as well as with the abundance of the main species and functional groups of the dung beetle assemblage (P < 0.01 or P < 0.05). Because of long-term agricultural cultivation, a lack of large herbivores and the recent use of larger paddocks than in the past, the species richness and abundance of functional group Ⅰ (i.e. larger paracoprids and telocoprids) were lower than in functional groups Ⅱ and Ⅲ. The functional group Ⅱ (relative smaller paracoprids) and Ⅲ (endocoprids) were the main components of the dung beetle assemblage in the study region. In summary, our results indicate the several methods of restoring cropland to grassland tested could promote the spatial heterogeneity of the farming landscape. Restoration of cropland to grassland, reasonable use of paddocks or using appropriate stocking levels when grazing livestock could be beneficial to enhancing and maintaining the species richness and abundance of functional group Ⅰ, as well as the biodiversity and overall ecological function of the dung beetle assemblage.