Abstract Fisheries bycatch of seabirds presents a serious management problem, and relatively little is known about this problem. In the Western North Atlantic, the issue is complicated by the relatively low effort and coverage of the observer programme and the high uncertainty in species identification. The Western North Atlantic is home to many endemic and endangered seabird populations, and the impact of fishery‐caused seabird bycatch has been of high interest, especially for those species with a low population size; however, species‐specific bycatch estimates have been difficult. From 1992 to 2016, 158 seabirds were observed caught by the US Atlantic pelagic longline fleet; among them, only 80 were identified to species, 25 were identified to family, and the rest – mainly in the older records – were unidentified. In this study, ecological traits of seabirds were used to improve bycatch estimation and provide species‐specific risk analysis to all the potentially affected seabird species in this region. Bayesian state–space modelling was used to accommodate the high level of uncertainty in the species identification process. Seabird bycatch risk was found to be highly dependent on population size. The group of large seabird species was estimated to be two times as vulnerable as the group of mid‐to‐small seabird species, scavenging and plunge‐diving feeding modes were identified as imposing high bycatch risks, and spatial and temporal distribution patterns were also good indicators of bycatch risk. Based on these ecological traits, shearwaters, gulls, gannets, and petrels were identified to potentially suffer from high bycatch in this region. These species, especially those that have not been identified historically, deserve extra attention in the observer programme, and for the implementation of conservation measures of seabirds in this region.