Soil erodibility is one of the key indicators of soil quality and is directly influenced by soil physical properties. This study aimed to Assessment soil physical quality indices and identify the most sensitive indicators across different soil erodibility classes in the semi-arid region of Zanjan Province, Iran. For this purpose, 20 soil physical properties were measured at 68 sampling points, including rangeland (45 points), cropland (13 points), orchard (8 points), and forest (2 points) land uses. Soils were classified into three erodibility classes based on the Williams erodibility factor. Differences among physical properties across erodibility classes were assessed using one-way ANOVA. Soil physical quality indices (SQI) were calculated using linear and nonlinear scoring functions with both simple and weighted approaches for the total dataset (TDS) and minimum dataset (MDS), and were compared using discriminant analysis. PCA identified three main components explaining 83% of the total variance. In the TDS, sand content, mean weight diameter of primary particles (MWDp), and mean weight diameter of dry aggregates (MWDdry) had the highest weights (W = 0.12, p < 0.01). In the MDS, sand content was the most important factor (W = 0.57, p < 0.01), followed by wet mean geometric diameter of aggregates (GMDwet) and MWDdry (W = 0.32 and 0.10, respectively). Discriminant analysis indicated that the best soil physical quality index for erodibility assessment was the weighted linear scoring method based on the TDS (SQI-2). Overall, soil structure stability and PSD were identified as effective tools for monitoring soil physical quality.