To evaluate the bottom-up NOX emission inventories with different data sources and spatial scales, an updated product of tropospheric NO2 vertical column densities (VCDs) from Ozone Monitoring Instrument (OMI),POMINO, was applied in chemistry transport modeling (CTM) and Gaussian function model for southern Jiangsu, a typical developed and polluted region in eastern China. Compared to the national emission inventory(MEIC), better correlation was found between spatial distributions of the high-resolution provincial inventory(JS) and POMINO VCDs. When applied in CTM, the simulated VCDs using JS were closer to POMINO data than those using MEIC, indicating the advantage of the provincial inventory that incorporated detailed information of individual plants. The simulated VCDs, however, were generally larger than observed ones, particularly for regions with high NO2 levels, partly because the improved NOX control measures for power sector were not fully considered in both national and provincial inventories. The “top-down” NOX emissions were estimated for four cities/city combinations in southern Jiangsu, using a Gaussian function model based on POMINO NO2 VCDs. The results were found to be most consistent with the estimates in JS among bottom-up inventories with different data sources. To further harmonize emissions and satellite observations at relatively small spatial scale, the online emission measurement data for individual plants are recommended for emission inventory development, and the products of satellite observation data with finer horizontal resolution are encouraged.