<p>Herpetofauna are particularly vulnerable to road mortality due to their movement behaviour and reliance on spatially discrete habitats. In tropical urban landscapes, such as university campuses, secondary forest patches act as critical refugia but are intersected by road networks that function as both barriers and mortality sources. This study examined the spatial clustering and drivers of herpetofauna road mortality within Universiti Malaysia Sabah (UMS). The study area was stratified into four zones based on traffic intensity, and monthly surveys were conducted along primary roads adjacent to forest edges. Roadkill incidents were georeferenced, while traffic volume and rainfall data were recorded. Kernel density estimation and Ripley’s K-function revealed significant clustering of roadkill events, with hotspots at a curved road and a junction in two zones adjacent to residential infrastructure. Roadkill frequency was higher during active academic semesters than during semester breaks, indicating a strong influence of human activity patterns. However, roadkill rate per vehicle was negatively correlated with traffic volume. This suggests that the risk of collision is higher during low-traffic periods, likely due to increased vehicle speeds or specific driver behaviours. These findings underscore the roles of road geometry, habitat configuration, and temporal patterns of human activity in shaping wildlife-vehicle collisions and highlight the need for targeted mitigation at identified hotspots.</p>