The current expansion of the road network is reaching unprecedented levels aimed to accelerate economic development and territorial expansion of human settlements, facilitating the transportation of people and goods. However, roads have several pervasive impacts on wildlife, such as pollution, habitat loss, roadkills or being barrier to animal movement. The most common mitigation action to minimize the last two impacts (roadkill and barrier effect) is the placement of wildlife barrier fencing and crossing structures. However, these public investments may not be completely effective, depending in part on the quality of their placement. To improve the ecological and cost-effectiveness of these mitigation actions, carefully choosing their locations is of paramount importance. Several methods have been employed to inform transportation agencies of the most suitable locations for wildlife fencing and crossings. From an ecological point of view, two contentious alternatives arise: a) to preserve movement areas; or b) to mitigate road sections with high rates of roadkills. We used GPS telemetry data from 90 mountain lions (Puma concolor) from California and a set of 74 roadkills from the same study area to explore similarities and dissimilarities between crossing points and roadkill hotspots to inform the best places for mitigation. The study area was in the coastal mountain ranges of Southern California, where mountain lions mainly occupy coastal shrub and forest ecoregions. This area includes megacities like Los Angeles and San Diego and their extensive metropolitan areas. The most common source of mortality (28% of deaths) of adult or subadult mountain lions in this region are vehicle collisions. We used GPS locations of 90 radio-collared mountain lions (43 females and 47 males; 40 adults, >30 months old at trapping, and 50 subadults) monitored in our study area between 2001–2022. We identified mountain lion road crossing locations (based on 5min intervals) by intersecting the movement paths with the road network and compared them to random locations in the road corridor. We selected only those locations with a road in a radius of 532m, because it represents the average displacement distance between each point over 5min time interval by mountain lions in our study area. We included as potential explanatory variables: terrain slope, distance to the nearest shrubland/scrubland, distance to the nearest forest, distance to the nearest grassland, distance to the nearest urban area and distance to the nearest water point, artificial light at night as illuminated landscapes, highway noise, and mule deer habitat suitability. Then, we compared our crossing points (i.e., safe crossings) to roadkill points by employing the same variables as above. Our results allow us to recommend the best places for mitigation as several crossing points were safe and were not likely to result in roadkills.