In some regions of the world, large carnivores, such as wolves, persist in landscapes with dense networks of paved roads. However, beyond the general impacts of roads on wildlife, we still lack information on carnivore responses to different types of roads and traffic volume levels. Using wolves in NW Spain as a case study, we show how wolves respond differently to paved road classes depending on road size, speed limit and traffic volume. All wolves evaluated (25 GPS collared wolves) crossed paved roads. Overall, during 3,915 sampling days, we recorded 29,859 wolf crossings. Wolf crossings of all paved road classes were recorded at a mean rate of 0.022 crossings/day/km (95% CI 0.016–0.027). Wolves crossed low speed and low traffic volume roads more frequently, and more often during the night, in order to lessen the chances of encountering traffic. We found mortality to be highest on roads with high speed and high traffic volume. How wolves interact with paved roads should be considered in landscape planning strategies in order to guarantee wolf long-term persistence in human-dominated landscapes. In our case, our results support an increasing focus on primary roads (class II) to identify segments of these roads where road mitigation efforts should be prioritised. Our study also highlights the importance of considering paved road classes when studying the impact of roads on wildlife.