Collisions with cars are an important mortality factor for many wild animals. Measures to mitigate road mortality are costly so should be implemented using cost-effective measures in locations where the road mortality is consistently highest and non-random in different species. It is thus important to identify what features causes these biodiversity collision blackspots. Almost all of the data and literature on collisions refer to vertebrates with little known about invertebrates. We used data on butterfly roadkills in three large landscape plots in Poland to identify sites where the collision rate seems to be routinely high. Biodiversity collision blackspots were identified from occurrence in successive years using spatial hierarchical clustering. Biodiversity collision blackspots comprised just 4% of the total road length, but included 49% of all road-killed butterflies. Habitats within 500 m of each blackspot was compared to random non-blackspot sites using generalized linear mixed models. The occurrence of blackspots was linked with high traffic volume, but only when cover of grassland in a landscape was high and verges had low plant species richness. Similarly, blackspots occurred with high probability when traffic volume was high but especially if grassland cover in the landscape and verge mowing frequency were also high. These blackspots had higher species richness and abundance of butterflies in the surrounding landscape than in random sites. Biodiversity collision blackspots analysis identified road sections of high road mortality for different butterfly species. Moreover, blackspots were also indication of species rich areas of conservation concern that were intersected by roads. Thus, conservation practitioners may direct mitigation measures, such as less frequent mowing and speed limit, in a cost-efficient manner in these spatially-limited locations.