The distribution of livestock across heterogeneous landscapes is often uneven, which has important implications for vegetation dynamics and how rangeland managers achieve desired outcomes from these landscapes. Here, we use data from widely available digital elevation models to classify a landscape in the shortgrass steppe with subtle topographic variation using two different approaches: topographic wetness index (TWI) and topographic position classes (TPCs) derived from topographic position indices. We used global positioning system collars to track the grazing locations of cattle within replicate pastures and fit generalized linear mixed models to their locations to quantify the influence of topography on grazing distribution. In addition, we examine the influence of the presence of saline vegetation communities on cattle use of lowlands. The resulting models indicate that TPC more effectively predicts grazing distribution than TWI and that the patterns are strongest in the second half of the growing season (August–October). Model performance was improved with the inclusion of saline vegetation communities, although the magnitude of cattle grazing time in these communities was not consistent across multiple pastures. These models, in combination with local knowledge, can be used by managers to predict and manage livestock distribution even in landscapes with relatively subtle topographic variability.