Application of fertilizers, unreasonable in rate and formula, has been the main factor limiting maize production. A large number of on-farm fertilization experiments have been conducted on maize across the main agro-ecological regions of China, and large volumes of first-hand data were collected for analysis in this study to explore region-specific optimal NPK recommendation (including rate and formula) in anticipation that the study may provide some technical support to improving the maize production in China and a theoretical basis for designing region-specific and crop-specific fertilizer. Based on the county’s GIS-based agricultural statistical database and relevant literature, the maize production areas in China were divided into 4 major regions by geographic location and production layout and further into 12 fertilization sub-regions in light of climatic, crop cultivation, topographic and soil conditions. Meanwhile a total of 1 752, 7 081 and 6 328 groups of on-farm N, P and K fertilization experiments were conducted all over the maize production areas during 2005-2010, and results and data of the experiments were collected, sorted and collated by fertilization sub-region. Then, sub-region-specific recommendation of N rate for maize was worked out using the N fertilizer total amount control technique and proper ratios of basal/sidedressing were determined in the light of the features and production practices of a region and nutrient demand of the crop; sub-region-specific recommendation of P rate for maize was, using P building-up and maintenance technique and sub-region-specific recommendation of K rate for maize, too, using the technique of yield response to K application. Based on the technical approach of ‘regional fertilizer formula plus site specific adjustment’, sub-region-specific fertilizer formula and application technique was designed for all the maize growing areas. Northeast China (RegionⅠ) and Northwest China (Region Ⅲ) was quite high in maize yield, reaching 9.0~10.5 t hm^-2, while North China Plain (RegionⅡ) and Southwest China (Region Ⅳ) was quite low, being 7.5~8.3 t hm^-2. For the 12 sub-regions, the recommended N rates were averaged to be N 181 kg hm^-2, ranging from N 150 kg hm^-2 for the cold sub-region in Northeast China (RegionⅠ-1) to N 219 kg hm^-2 for the irrigated oasis sub-region in the Northwest China (RegionⅢ-3); the recommended P rates, P₂O₅ 75 kg hm^-2 on average, ranged from P₂O₅ 46 kg hm-2 for the north-central sub-region in the North China Plain (RegionⅡ-1) to P₂O₅ 123 kg hm^-2 in the irrigated oasis sub-region in the Northwest China (RegionⅢ-3); and the recommended K rates, K₂O 54 kg hm^-2 on average, ranged from K₂O 30 kg hm^-2 for the irrigated oasis sub-region in the Northwest China (RegionⅢ-3) to K₂O 64 kg hm^-2 for the south sub-region in the North China Plain (RegionⅡ-2). For the 12 sub-regions, a total of 16 region-specific fertilizer formulas were worked out, including 12 formulae with basal plus top-dressing fertilization approach and 4 formulae with only basal fertilization application. The regionalization of maize fertilization can well reflect the layout of maize production in China and region-specific climatic, topographic, soil and relevant crop cultivation conditions. The significant differences between the agro-ecological regions in maize grain yield and recommended fertilizer rates reflected the specificity of each region in soil fertility, yield response to nutrient application and crop nutrient demand. The region-specific fertilizer formulas designed on the basis of the region-specific fertilizer recommendation and characteristics of the regions well match the maize productions in the regions and optimize regional fertilizer application techniques. These region-specific fertilizer formulas will play an important role in guiding optimization of compound fertilizer products and provide a theoretical reference for designing regional fertilizer formulas for relevant crops in future.