Objective  The aims of the study were to obtain the soil nutrient plentiful-lack index, to establish nutrient classification index and recommended fertilization, to guide potato scientific fertilization, and to improve the fertilizer utilization rate of potatoes in the semi-arid area of central Gansu.

  Method  By analyzing the potato “3414” test data in Huining county from 2015 to 2018, the correlation between soil available nutrient content and potato relative yield and economic optimal fertilization was studied, the soil nutrient classification index and recommended fertilization were established in different classification ranges.

  Result  The correlation between soil total nitrogen (N) content and yield and N economic and optimum fertilization was better than soil alkali-hydro N in the potato field, the total N, available phosphorus (P), and available potassium (K) were determined as abundant and deficient indices, and recommended fertilization calculation indices on potato fields in Semi-arid Area of Central Gansu. The fertilizer interval regulation of nutrient abundant and deficient indices was established, including the quantitative micro regulation model of target yield and fertilizer amount. The total N content of potato soil in the study area was divided into five fertility grades, i.e., low (< 0.6 g kg⁻¹), lower (0.6-1.1 g kg⁻¹), medium (1.1-1.8 g kg⁻¹), higher (1.8-2.2 g kg⁻¹) and high (> 2.2 g kg⁻¹); The available P was divided into four fertility grades, i.e., lower (< 13 mg kg⁻¹), medium (13-25 mg kg⁻¹), higher (25-31 mg kg⁻¹) and high (> 31 mg kg⁻¹); The available K was divided into four fertility grades, i.e., lower (< 162 mg kg⁻¹), medium (162-238 mg kg⁻¹), higher (238-271 mg kg⁻¹) and high (> 271 mg kg⁻¹).

  Conclusion  Based on the fertilizer effect function and the average target yield of potatoes, the recommended fertilization rates in different fertility levels were put forward. According to the soil fertility grade from low to high, the fertilization rates are N > 158.0, 134.5-158, 113.7-134.5, 105.9-113.7 and 0-105.9 kg hm⁻², P₂O₅ > 136.7, 90.4-136.7, 75.7-90.4 and 0-75.7 kg hm⁻², K₂O > 82.2, 56.8-82.2, 47.0-56.8 and 0-47.0 kg hm⁻², respectively. In production, the economic and optimal amount of fertilizer could be output by detecting the soil nutrients of any plot and substituting them into the correlation function model.