Objectives Evaluation of soil health is an effective measurement for identifying the soil health status and main soil obstacle factors of melon fields, which provides important theoretical support for the cultivation of healthy soil. This study established the soil health assessment database and evaluated the soil health index of melon fields by constructing a minimum data set (MDS) in Turpan City, Xinjiang.

Methods A total of 72 soil samples were collected in the tillage layer of melon fields in Tuyugou, Ya’er, and Sanbao towns, Turpan City in Xinjiang, all the tested fields had planted melon for averaged 5 years. Twenty-eight soil physical, chemical, and biological properties were analyzed as the health indicators. A soil health assessment database was established, and the minimum data set (MDS) was constructed using principal component analysis. Linear and non-linear soil health assessment methods were used for evaluation. The total data set (TDS) was used for verification.

Results Seven indicators were selected for the MDS by the principal component analysis, including the soil bulk density, electrical conductivity, available phosphorus, organic carbon, exchangeable calcium, soil respiration, and β-glucosidase. For the linear scoring function, the soil health indices calculated based on the TDS and MDS were 0.44 and 0.49, respectively, with coefficients of variation of 22.5% and 18.7% correspondingly. For the non-linear scoring function, the average soil health indices calculated based on the TDS and MDS were 0.46 and 0.47, respectively, while the coefficients of variation were 25.8% and 24.3% in that order. Using linear and non-linear evaluation methods, there was a significant positive correlation between the soil health index calculated based on the TDS and MDS (P<0.01). The slope is close to 0.9. Overall, the soil health index of Tuyugou and Ya’er Town was higher than that of Sanbao Town.

Conclusions The MDS can be used as a substitute for the TDS in assessing soil health in melon fields. The large coefficient of variation of the soil health indices calculated by the nonlinear scoring function indicates that the nonlinear scoring function is more sensitive. Therefore, it is recommended to select the nonlinear scoring function for soil health evaluation of melon fields. The melon field in Turpan City, Xinjiang, is alkaline in terms of pH, with rich soil nutrient content, but relatively low organic matter and active carbon. According to the assessment of MDS, the soil health index in Turpan is bellow 0.5, indicating a moderate level of health.