Objectives This study was to optimize the nitrogen, phosphorus and potassium fertilizer rates and placement for high yield of hybrid millet, and to optimize fertilization scheme for providing foundation of reducing fertilization rate, increasing efficiency and maintaining sustainable development in hybrid millet production.

Methods A field experiment was conducted during 2016 to 2017 at the Innovation Zone of Shanxi Agricultural University. Quadratic regression orthogonal rotation combination design was used in the experiment; hybrid millet (Setaria italica) cultivar of Zhangzagu10 was used as the tested material and cultivated with plastic film mulching. It included 5 factors as nitrogen, phosphorus, potassium fertilizer rates, the horizontal distance to seed and depth of application, each factor had 5 levels. All the fertilizers were basal applied in the designed rate and position. At maturity, 10 clumps of millet plants were taken to examine 1000-grain weight, grain number per ear and yield. The correlation between the tested index with the fertilizer rate and placement were established using multivariate linear regression model.

Results Millet yield was extremely significantly dependent on grains per ear and 1000-grain weight, and significantly on the ears per hectare. Fertilizer rates and placement had extremely significant effects on the yield, and the effect descended in order of nitrogen > phosphorus > potassium >fertilization horizontal distance to seed> fertilization depth. Nitrogen had a significant effect on grains per ear; phosphorus had an extremely significant effect on ears per hectare and had a significant effect on 1000-grain weight; potassium had an extremely significant effect on ears per hectare and grains per ear and had a significant effect on 1000-grain weight; and the fertilization horizontal distance and depth had significant effects on ears per hectare and grains per ear. The interaction of N and P was significant on yield, that of N and K was significant on grains per ear, 1000-grain weight and yield; and that of P and K was significant on ears per hectare. The regression among the five factors with yields and yield components fitting well, showing prospect in actual prediction.

Conclusions For hybrid millet cultivar of Zhangzagu10, the rates of NPK fertilizer affect more than position on the yield. N increases grain number per ear, P increases ear numbers per hectar and 1000-grain weight, and K increases all above. Under the tested conditions, the optimal fertilizer rate is N 225 kg/hm², P₂O₅ 136–154 kg/hm² and K₂O 94–109 kg/hm². The suitable horizontal distance to seed and depth of fertilization is 17–19 cm and 21–24 cm.