Purpose This study aimed to quantify the individual and interactive effects of rainfall intensity, slope gradient, and soil texture on rill morphological characteristics, and determine the dominant factor controlling rill erosion processes on the Loess Plateau of China. Methods Indoor simulation experiments were conducted, testing two rainfall intensities (90 and 120 mm h-& sup1;), three slope gradients (10 degrees, 15 degrees, 20 degrees) and two soils from the south (loamy clay, LC) and north (sandy loam, SL) parts of the Loess Plateau. High-resolution 3D laser scanning was used to quantify both basic and derived rill morphological indicators. The individual and interactive effects of the factors were analyzed using correlation analysis and the GeoDetector model. Results Rill morphology was jointly controlled by all three factors, with slope gradient being the dominant influence. Increasing slope gradient significantly enhanced rill density (rho), contour deviation (& micro;), and degree of cleavage (delta), indicating the higher degree of rill development and surface fragmentation on steeper slopes. While average width-to-depth ratio (RWDa) and shape ratio (etaa) deceased in steeper slopes, reflecting a transition from wide-shallow to narrow-deep rill cross-sections with the rising slope gradient. Higher rainfall intensity primarily promoted lateral erosion, whereas the clay-rich LC soil was more susceptible to vertical downward cutting erosion compared to SL. Pairwise factor interactions showed stronger effects than individual factors, among which the slope gradient-soil texture interaction resulted in the pronounced non-linear enhancement on both cross-sectional and network indicators. Conclusion Slope gradient governs rill development by controlling runoff energy and regulating the effects of rainfall intensity and soil texture. These findings deepen the mechanistic understanding of rill erosion processes and highlight the necessity of incorporating multi-factor interactions into rill erosion simulation models to achieve more accurate soil loss prediction and effective risk assessment for hillslope croplands on the Loess Plateau.