Excessive or imbalanced fertilization has not only decreased nutrient use efficiency, but also degraded arable land and posed a great threat to the environment. In this study, the datasets were collected from field experiments for the period 1992 to 2017 in the main potato (Solanum tuberosum L.) production regions of China. We used the Quantitative Evaluation of the Fertility of Tropical Soils (QUEFTS) model to estimate the soil‐plant balanced N, P, and K requirements for the potato in China. Our results revealed that there were great differences in the potato yield and nutrient uptake. The upper and lower 2.5th percentiles of N, P, and K internal efficiencies (IE, kg tuber per kg nutrient in the total plant) data were used as maximum accumulation (a) and maximum dilution (d) boundary parameters in the QUEFTS model, which were 133 and 463 kg kg–1 for N, 652 and 3030 kg kg–1 for P, and 119 and 790 kg kg–1 for K, respectively. The QUEFTS model predicted plant nutrient uptake of 4.0 kg N, 0.7 kg P, and 3.5 kg K to produce 1 Mg of tuber in the linearly distributed portion, and a corresponding tuber nutrient demand of 2.9 kg N, 0.5 kg P, and 2.3 kg K to produce 1 Mg of tuber. Field validation experiments confirmed that the QUEFTS model could be used to simulate optimum nutrient uptake and provide appropriate parameters in building fertilizer recommendation, which helps improve nutrient use efficiency for potato in China.Core Ideas This is the first report of QUEFTS model on estimating the optimum nutrient requirements for potato crops in China. The datasets used in this study were collected from multi‐year‐site field experiments. Multi‐site field validation experiments in potato‐producing areas confirmed the feasibility of QUEFTS model‐simulated nutrient uptake.