The northwestern Barents Sea (NW-BS) is a highly productive region within the transitional zones of an Atlantic to Arctic-dominated marine ecosystem. The steep latitudinal gradients in sea ice concentration, Atlantic and Arctic Water, offer an opportunity to test hypotheses on physical drivers of spatial and temporal variability of net primary production (NPP). However, quantifying NPP in such a large ocean region can be challenging by the lack of in situ measurements with high spatial and temporal resolution, and gaps in remote sensing estimates due to the presence of clouds and sea ice, and assumptions regarding the depth distribution of biomass. Without reliable data to evaluate models, filling these gaps with numerical models is limited by the model representation of the physical environment and its assumptions about the relationships between NPP and its main limiting factors. Hence, within the framework of the Nansen Legacy Project, we combined in situ measurements, remote sensing, and model simulations to constrain the estimates of phytoplankton NPP in NW-BS. The region was subdivided into Atlantic, Subarctic, and Arctic subregions on the basis of different phytoplankton phenology. In 2004 there was a significant regime change in the Atlantic subregion that resulted in a step-increase in NPP in tandem with a step-decrease in sea ice concentration. However, neither region experienced significant long term trends in NPP despite changes in the physical environment. Mixing was the main driver of simulated annual NPP in the Atlantic subregion, while light and nutrients drove annual NPP in the Subarctic and Arctic subregions. The multi-source estimate of annual NPP ranged 79–118 gC m−2 yr−1 in the Atlantic, 74–82 gC m−2 yr−1 in the Subarctic, and 19–47 gC m−2 yr−1 in the Arctic. The total NPP in the NW-BS region was estimated between 15 and 48 Tg C yr−1, which is 15–50% of the total NPP needed to sustain three of the most harvested fish species north of 62°N (roughly  90 Tg C yr−1). This research shows the importance of continuing to strive for better regional estimates of NPP.