It is widely recognized that major transportation infrastructures such as highways degrade habitats, increase the risk of collision mortality, and cause alterations in landscape connectivity by reducing connections between habitats due to the barrier effect. Wildlife crossings (underpasses and overpasses) have therefore become standard practice to reduce these effects on wildlife. Most studies aim to identify relevant locations for the creation of new wildlife crossings, but it is very rare for results based on modeling tools or animal movement records, for example, to be confronted with expert opinion and knowledge. Our paper provides a methodological framework combining modeling tools and field expertise from a multi-scale perspective. We used complementary connectivity metrics from graph theory to accurately quantify and spatialize the most relevant areas to improve habitat connectivity with multispecies conservation goals. The connectivity analysis was coupled with field expertise to take into account the local context, and technical and land feasibility. We prioritized existing road structures for improvement, found the best locations for new wildlife crossings, and made specific, costed technical recommendations. We empirically tested this approach on a large highway network in Northern France. This allowed us to define work principles to improve 109 existing road structures and to create underpasses or overpasses on 15 road sections. This innovative study for operational purposes has been a useful tool to assist the decision-making process of land-use planners who wish to improve wildlife crossings on major transportation infrastructures.