The Northwest Pacific (NWP) is an important source region of nitrous oxide (N₂O) in the global ocean, yet its spatiotemporal variability and its underlying drivers remain poorly constrained due to the scarcity of observations in this key region. Here, based on high-resolution underway N₂O observations from five cruises conducted in the NWP, we developed a machine learning approach to reconstruct the spatiotemporal variability of surface-water ΔN₂O and evaluate the corresponding air–sea N₂O flux in this key region. The reconstructed regional N₂O flux was estimated to be 171.97 ± 4.95 Gg N yr⁻¹, accounting for approximately 3.9–6.7% of global oceanic N₂O emission. N₂O flux exhibited pronounced seasonal and spatial heterogeneity, the maximum emission occurred in winter and the minimum was in summer. Seasonal variability was primarily regulated by changes in gas transfer velocity (Kw), while variations in surface-water ΔN₂O was important in specific regions. Spatially, N₂O fluxes generally increased toward higher latitudes, reflecting the combined effects of vertical mixing, lateral transport, and biogeochemical processes across different oceanic regimes. Furthermore, the reconstructed N₂O emissions in the Northwest Pacific exhibited a declining trend during 2015–2024, decreasing at a rate of −1.22 Gg N yr⁻¹. This study reveals the multiscale variability and controlling mechanisms of N₂O emissions in the Northwest Pacific and provides new observational constraints for assessing regional oceanic N₂O sources and their potential climate impacts.