Résumé: Trawling activities are considered to be one of the main sources of disturbance to the seabed worldwide. We aimed to disentangle the dominance of environmental variations and trawling intensity in order to explain the distribution of diversity patterns over 152 sampling sites in the French trawl fishing-ground, the Grande Vasière. Using a towed underwater video device, we identified 39 taxa to the finest taxonomic level possible, which were clustered according to their vulnerability to trawling disturbance based on functional traits. Using generalized linear models, we investigated whether the density distribution of each vulnerability group was sensitive to trawling intensity and habitat characteristics. Our analyses revealed a structuring effect of depth and substratum on community structure. The distribution of the more vulnerable group was a negative function of trawling intensity, while the distributions of the less vulnerable groups were independent of trawling intensity. Video monitoring coupled with trait-based vulnerability assessment of macro-epibenthic communities might be more relevant than the traditional taxonomic approach to identifying the areas that are most vulnerable to fishing activities in conservation planning.