The vulnerability of grassland ecosystems in the Mongolian Plateau has increased due to climate change and human intervention, and thus herdsmen must adopt additional livelihood adaptation strategies. In this study, we determined the impact of the livestock structure, that is, the proportion of sheep and goats in the livestock population, on grassland coverage on the Mongolian Plateau (Mongolia and the Nei Mongolia Autonomous Region of China). By applying a fixed effect model and mediating effects model to a unique combination of statistical and remote sensing data for 20 border provinces/banners from 2007 to 2016, we found that a higher proportion of sheep and goats led to grassland degradation, and the grassland type (i.e., whether among the top 50% in terms of the Normalized Difference Vegetation Index) could alleviate the negative impact of a high proportion of sheep or goats on grassland coverage. The results obtained by the mediating effects model showed that the primary industry gross domestic product (GDP) had a masking effect, where it could mitigate the negative influence of sheep or goats on grassland coverage. Regional heterogeneous analysis did not provide significant results for Mongolia due to its relatively stable livestock structure, but the livestock structure in Nei Mongolia had a greater negative impact on grassland coverage, thereby further indicating that a decrease in the proportion of sheep and goats had a greater ecological benefit.