Estimating the feed intake of grazing herbivores is critical for determining their nutrition, overall productivity and utilization of grassland resources. A 17-day indoor feeding experiment was conducted to evaluate the potential use of Medicago sativa as a natural supplement for estimating the total feed intake of sheep. A total of 16 sheep were randomly assigned to four diets (four sheep per diet) containing a known amount of M. sativa together with up to seven forages common to typical steppes. The diets were: diet 1, M. sativa + Leymus chinensis + Puccinellia distans; diet 2, species in diet 1 + Phragmites australis; diet 3, species in diet 2 + Chenopodium album + Elymus sibiricus; and diet 4, species in diet 3 + Artemisia scoparia + Artemisia tanacetifolia. After faecal marker concentrations were corrected by individual sheep recovery, treatment mean recovery or overall recovery, the proportions of M. sativa and other dietary forages were estimated from a combination of alkanes and long-chain alcohols using a least-square procedure. Total intake was the ratio of the known intake of M. sativa to its estimated dietary proportion. Each dietary component intake was obtained using total intake and the corresponding dietary proportions. The estimated values were compared with actual values to assess the estimation accuracy. The results showed that M. sativa exhibited a distinguishable marker pattern in comparison to the other dietary forage species. The accuracy of the dietary composition estimates was significantly (P < 0.001) affected by both diet diversity and the faecal recovery method. The proportion of M. sativa and total intake across all diets could be accurately estimated using the individual sheep or the treatment mean recovery methods. The largest differences between the estimated and observed total intake were 2.6 g and 19.2 g, respectively, representing only 0.4% and 2.6% of the total intake. However, they were significantly (P < 0.05) biased for most diets when using the overall recovery method. Due to the difficulty in obtaining individual sheep recovery under field conditions, treatment mean recovery is recommended. This study suggests that M. sativa, a natural roughage instead of a labelled concentrate, can be utilized as a dietary supplement to accurately estimate the total feed intake of sheep indoors and further indicates that it has potential to be used in steppe grassland of northern China, where the marker patterns of M. sativa differ markedly from commonly occurring plant species.