Clarifying the effects of different vegetation types on soil quality is of great significance for maintaining ecological stability in the hilly-gully region of the Loess Plateau. This study was conducted in the Anjiapo watershed， Dingxi City， Gansu Province as an example. Five vegetation types were selected， including spruce forestland， sea-buckthorn forestland， alfalfa grassland， wheat cropland and abandoned bare land. A total of 18 soil indicators were measured in the 0-20 cm soil layer. Soil quality indices were calculated based on the total data set， minimum data set， and revised minimum data set， combined with linear and non-linear scoring methods. The results showed that spruce forestland had the highest SQI of 0.53， which was 35.9% higher than that of abandoned bare land. The revised minimum data set combined with the linear scoring model （SQI-LRM） exhibited the strongest explanatory power. Partial least squares structural equation modeling revealed that different vegetation types primarily affected soil biological properties through their influence on soil chemical properties. Random forest combined with multiple linear regression analysis revealed that different vegetation types mainly drove soil quality by regulating soil pH， which in turn influenced protease activity. In conclusion， spruce forestland can significantly improve soil quality， and the SQI-LRM model can provide a quantitative basis for soil quality improvement and sustainable ecosystem development in the study area.

Loess Plateau  /  Different vegetation types  /  Revised minimum dataset  /  Soil quality assessment  /  Scoring method