Abstract: To investigate an approach of provincial GDP spatialization at pixel level, using NPP-VIIRS nighttime light image as data source, isolated exceeding-bright pixels and background noises were firstly filtered.  Then total nighttime light (TNL), linear weighted light index (L) and compounded night light index (CNLI)  were respectively calculated and analyzed on correlation with all counties′ GDP including in each prefecture-level city of Hebei province. TNL and all counties′ GDP included in each prefecture-level city showed the strongest correlation in three light indexes. The optimum regression model corresponding to TNL in each prefecture-level city was used to calculate each pixel's GDP. Once these GDPs at pixel level were linearly calibrated, a Hebei province′s GDP density map was generated based on these linearly corrected GDPs. The results showed that relative error of estimated average GDP′s relative error was 0.10% among 172 counties of Hebei province. Additionally, this approach had higher precision and produced GDP density map could reflect the current economic development situation of Hebei province.