The monodactyl horses of the genus Equus originated in North America during the Pliocene, and from the beginning of the Pleistocene, they have been an essential part of the large ungulate communities of Europe, North America and Africa. Understanding how body size of Equus species evolved and varied in relation to changes in environments and diet thus forms an important part of understanding the dynamics of ungulate body size variation in relation to Pleistocene paleoenvironmental changes. Here we test previously published body mass estimation equations for the family Equidae by investigating how accurately different skeletal and dental measurements estimate the mean body mass (and body mass range) reported for extant Grevy's zebra ( Equus grevyi) and Burchell's zebra ( Equus quagga). Based on these tests and information on how frequently skeletal elements occur in the fossil record, we construct a hierarchy of best practices for the selection of body mass estimation equations in Equus. As a case study, we explore body size variation in Pleistocene European Equus paleopopulations in relation to diet and vegetation structure in their paleoenvironments. We show a relationship between diet and body size in Equus: very large-sized species tend to have more browse-dominated diets than small and medium-sized species, and paleovegetation proxies indicate on average more open and grass-rich paleoenvironments for small-sized, grazing species of Equus. When more than one species of Equus co-occur sympatrically, the larger species tend to be less abundant and have more browse-dominated diets than the smaller species. We suggest that body size variation in Pleistocene Equus was driven by a combined effect of resource quality and availability, partitioning of habitats and resources between species, and the effect of environmental openness and group size on the body size of individuals.