Herbivores regulate nutrient cycling of terrestrial ecosystems through trampling topsoil and vegetation, selective foraging, and excretion of feces and urine. However, the role of trampling in the soil-plant interaction is still unclear, partly due to the lack of empirical studies examining soil and plant stoichiometric responses to trampling. We conducted a 2-consecutive-yr field trial to explore the effects of simulated sheep trampling intensity on soil and plant carbon:nitrogen:phosphorus (C:N:P) stoichiometry in a typical steppe of the Loess Plateau, China. Results show that with the increase in trampling intensity, the soil bulk density at 0–10 cm decreased in 2016 while it increased in 2017. Although trampling increased soil total N and P concentration, the soil C:N, C:P, and N:P ratios remained stable. The aboveground biomass of three dominant species increased with trampling intensity in 2016 but decreased in 2017. Trampling increased plant N and P concentrations but decreased plant C concentration, and C:N and C:P ratios. Trampling affected plant N:P ratios depending on species, but usually exacerbated P limitation. Thus, the differential responses of soil and plant C:N:P stoichiometry to trampling intensity indicated that the effects of trampling on the plants did not convey equivalent impacts on the soil. Our study provides evidence that it is necessary to isolate the effects of livestock trampling from grazing on grasslands and highlights that at the stocking rate of 2.7 Tan sheep ha^–1 (equivalent to ∼40 footsteps m^–2), trampling does not negatively affect soil-plant interactions or stoichiometry, and hence such stocking rate may be compatible with rangeland restoration objectives.