The mortality of wildlife due to run over is one of the main threats to biodiversity. Several factors may be related to the occurrence of such roadworks, including the intense traffic of vehicles, the recklessness of drivers, the displacement of animals crossing the highways, as well as the landscape patterns. The present study aimed at identifying points of greatest trampling of wild mammals on roads surrounding the Seridó Ecological Station (ESEC Seridó) and their relationship with elements of the surrounding landscape. In addition, the relationship between run-over and temporal variation was evaluated. The selected sections were monitored at 21-day intervals for 36 months. For the spatial approach, the data were obtained from the classification and quantification of the landscape units using 1 km, 5 km and 10 km lightning buffers using two Landsat images (one from the rainy season and one from the rainy season dry), where it was classified as the presence of bodies of water, native vegetation, agriculture, exposed soil and urbanization. For the temporal characterization, the variables of temperature, rainfall, humidity and photoperiod were used. Pearson's correlation analysis was used to detect significant relations and Akaike's tests to identify the most explanatory models about relations. A higher rate of run-of-the-road mammals was observed during the dry season. The species of wild mammals most affected by runners were Cerdocyon thous, Lycalopex vetulus and Procyon cancrivorus. Spatially, Mammalia Class hotspots were negatively related to agriculture and urbanization, while Cerdocyon thous hotspots were positively related to native vegetation and negatively related to agriculture, exposed soil and urbanization, and Procyon cancrivorus hotspots, by their positively related to urbanization. The Lycalopex vetulus species had no significant relationship. Regarding the temporal variables, the general group of mammals and the species Cerdocyon thous presented negative relation to the photoperiod