Abstract: The GDP data in China are available in the format of statistical tables with the spatial scale of administrative district in GIS. Sometimes they are not instable, imprecise and not unified with most geographic datasets. GDP in pixels instead of administrative polygons will facilitate the integration and comprehensive analysis between GDP and remote sensing products, such as land use and land cover changes. GDP datasets of China are composed of primary industry GDP, secondary industry GDP, and tertiary industry GDP. The correlative relationships between the land use and three industries were analyzed, and the spatial correlation model was implemented. The model was used to calculate the statistical GDP of every county for each pixel. The 1 km grid GDP datasets of China in 2005 and 2010 were derived. Forty counties including township GDP data were selected for precision verification. The results showed that the errors of the spatial GDP data in these counties were between 6% and 17%.

Key words: China, GDP, grid, spatial data, 2005, 2010