The experiment was conducted in the same field in Changzhi, Shanxi for two consecutive years, the experimental design adopted the "3414" scheme in the quadratic regression d-optimum design, under exploring the effects of different nitrogen（N）, phosphorus（P） and potassium（K） application amount on the fertilizer efficiency, yield and economic benefit of millet. The results showed that, both the ternary quadratic and unitary quadratic models for fertilizer application in this experiment were unsuccessful. The application of P fertilizer at Level 2（60 kg/ha） was found to be unfavorable for increasing millet yield. The treatment on N₃P₂K₂（N 120 kg/ha, P₂O₅ 60 kg/ha, K₂O 90kg/ha）cost the most, however, it’s economic income decreased by 2,194.19 yuan/hacompared with the control. The treatment on N₂P₁K₁（N 90 kg/ha, P₂O₅ 30 kg/ha, K₂O 45 kg/ha）was the most beneficial to increase yield and income of millet. The potential factors contributing to the unsuccessful fit of the "3414" fertilizer application model could be attributed to unreasonable fertilizer application, including the excessive application of P fertilizer at Level 2, while the appropriate application rates of N and K fertilizers require further investigation. Therefore, it is recommended to control the amount of fertilizer applied within a reasonable range during millet cultivation, and to avoid blindly investing excessive labor and resources in the pursuit of high yields. It is essential to be realistic and achieve optimal benefits by considering practical factors.