Seabirds are vulnerable to incidental harm from human activities in the ocean, and knowledge of their seasonal distribution is required to assess risk and effectively inform marine conservation planning. Significant hydrocarbon discoveries and exploration licenses in the Labrador Sea underscore the need for quantitative information on seabird seasonal distribution and abundance, as this region is known to provide important habitat for seabirds year-round. We explore the utility of density surface modeling (DSM) to improve seabird information available for regional conservation and management decision making. We, (1) develop seasonal density surface models for seabirds in the Labrador Sea using data from vessel-based surveys (2006-2014; 13,783 linear km of surveys), (2) present measures of uncertainty in model predictions, (3) discuss how density surface models can inform conservation and management decision making, and 4) explore challenges and potential pitfalls associated with using these modeling procedures. Models predicted large areas of high seabird density in fall over continental shelf waters (max. ∼80 birds·km −2 ) driven largely by the southward migration of murres (Uria spp.) and dovekies (Alle alle) from Arctic breeding colonies. The continental shelf break was also highlighted as an important habitat feature, with predictions of high seabird densities particularly during summer (max. ∼70 birds·km −2 ). Notable concentrations of seabirds overlapped with several significant hydrocarbon discoveries on the continental shelf and large areas in the vicinity of the southern shelf break, which are in the early stages of exploration. Some, but not all, areas of high seabird density were within current Ecologically and Biologically Significant Area (EBSA) boundaries. Building predictive spatial models required knowledge of Distance Sampling and GAMs, and significant investments of time and computational power-resource needs that are becoming more common in ecological modeling. Visualization of predictions and their uncertainty needed to be considered for appropriate interpretation by end users. Model uncertainty tended to be greater where survey effort was limited or where predictor covariates exceeded the range of those observed. Predictive spatial models proved useful in generating defensible estimates of seabird densities in many areas of interest to the oil and gas industry in the Labrador Sea, and will have continued use in marine risk assessments and spatial planning activities in the region and beyond.