The subfamily Equinae in the Great Plains region of North America underwent a dramatic radiation and subsequent decline as climate changed from warm and humid in the middle Miocene to cooler and more arid conditions during the late Miocene. Here we use ecological niche modeling (ENM), specifically the GARP (Genetic Algorithm using Rule-set Prediction) modeling system, to reconstruct the geographic distribution of individual species during two time slices from the middle Miocene through early Pliocene. This method combines known species occurrence points with environmental parameters inferred from sedimentological variables to model each species' fundamental niche. The geographic range of each species is then predicted to occupy the geographic area within the study region wherever the set of environmental parameters that constrain the fundamental niche occurs. We analyze changes in the predicted distributions of individual species between time slices in relation to Miocene/Pliocene climate change. Specifically, we examine and compare distribution patterns for two time slices that span the period from the mid-Miocene (Barstovian) Climatic Optimum into the early Pliocene (Blancan) to determine whether habitat fragmentation led to speciation within the clade and whether species survival was related to geographic range size. Patchy geographic distributions were more common in the middle Miocene when speciation rates were high. During the late Miocene, when speciation rates were lower, continuous geographic ranges were more common. Equid species tracked their preferred habitat within the Great Plains region as well as regionally throughout North America. Species with larger predicted ranges preferentially survived the initial cooling event better than species with small geographic ranges. As climate continued to deteriorate in the late Miocene, however, range size became irrelevant to survival, and extinction rates increased for species of all range sizes. This is the first use of ENM and GARP in the continental fossil record. This powerful quantitative biogeographic method offers great promise for studies of other taxa and geologic intervals.