Identifying air pollutant sources and quantifying related emissions provides crucial information towards improving global air quality and public health. Emission inventories commonly used for oil and gas (OG) activities inadequately represent nitrogen oxide (NOx) emissions, with notable discrepancies identified in the past. Satellite remote sensing provides a unique vantage point to map and quantify several key multi-pollutant species on a routine basis across the global scale. Here, we quantify annual NOx emissions from 44 major OG basins distributed globally, utilizing TROPOspheric Monitoring Instrument (TROPOMI) nitrogen dioxide (NO2) observations with the divergence flux method. In addition, we use the spaceborne Visible Infrared Imaging Radiometer Suite (VIIRS) natural gas flaring detections to further constrain satellite-derived NOx emissions. The divergence flux method, which addresses 3D topography corrections and chemical loss of NOx while accounting for wind-induced flux smearing, provides a robust approach for estimating NOx emissions. Our findings reveal a substantial level of NOx emissions from individual OG facilities and major production clusters as well as flaring activity, which serve as valuable data for assessing emission inventories. Subsequently, we conducted a comparative analysis between our TROPOMI-derived NOx emissions and those obtained from the existing EDGARv6.1 and CAMS-GLOB-ANT_v5.3 global emission inventories. This evaluation underscored a noteworthy disparity, with existing inventories significantly reporting lower emissions from the OG sector relative to TROPOMIderived data. Specifically, EDGAR exhibited a substantially lower estimate of onshore emissions by 61%, and CAMS displayed an even pronounced lower estimate by 78%, compared to TROPOMI-derived data from this study. For offshore emissions, EDGAR showed a similar pattern, albeit to a lesser extent, with a 26% underestimation, while CAMS showed a considerable underestimation of 92%. These findings hold significance for enhancing global NO2 emission inventories, particularly from OG fields, and therefore address gaps in current bottom-up inventories. Additionally, our approach yields detailed spatial emission maps, enhancing granularity in depicting NOx distribution. We further explore the variations in the correlation between NOx emissions with OG production volumes and CH4 concentrations across the different OG basins in North America, shedding light on their emissions characteristics and providing insights into the observed co-locations of methane and NOx emissions from oil and gas activity.