The spatial and temporal incidence of possum mortality on roads was monitored in Sydney’s northern suburbs over a two-year period. In total, 217 road surveys were completed, equating to over 17 000 km travelled. Almost 600 possums were observed as road-kill, with common ringtail possums (Pseudocheirus peregrinus) encountered in 87% of surveys and almost three times as often as common brushtail possums (Trichosurus vulpecula) (encountered in 52% of surveys). Roads were classified into three categories based on adjacent landscape attributes, with the road-side environment ranging from suburban housing to continuous sclerophyll forest. Clusters, or ‘hot spots’, for possum road-kill were identified using GIS. Most possum fatalities, and five out of six hot spots, occurred along roads at the bush–urban interface, suggesting high possum movement rates and/or abundance at these sites. Continuous canopy across the road was available along only 4% of the roads surveyed, forcing these arboreal marsupials to cross roads via electricity wires or at ground level, making them vulnerable to collision with vehicles. In an attempt to mitigate the possum road toll, two rope tunnel bridges were erected at canopy level above the most consistent hot spot. Subsequent monitoring of the roads was undertaken for 10 months after erection of the canopy bridge and road-kill numbers declined significantly at the site of the bridges, but also at adjacent sites throughout the study area.