Although vegetation is widely recognised for its ability to reduce slope erosion, the specific characteristics of changes in rill morphology parameters on grass cover (Vc) slopes remain incompletely understood. Furthermore, accurately predicting soil erosion rates on Vc slopes is crucial for assessing soil erosion and developing predictive models. This study investigated the variation characteristics of erosion and rill morphology parameters on Vc slopes through indoor artificial rainfall simulation experiments. The results indicate: (1) Vc markedly suppresses slope erosion; relative to bare slopes, mean runoff rates on Vc slopes decreased by 20.70%-35.61%, and mean erosion rates decreased by 30.28%-90.50%. (2) The width, depth, density, fractal dimension and fragmentation of rills all exhibited a decreasing trend with increasing Vc. Conversely, the width-to-depth ratio of rills initially decreased then increased with increasing Vc, with this transition occurring within the 30%-40% Vc range. (3) Among the morphological parameters, rill density best characterises soil erosion in this study. Furthermore, the slope erosion rate prediction equation developed using rainfall intensity, slope gradient, Vc and rill density demonstrated high accuracy and predictive efficacy. The findings may provide useful insights for constructing soil erosion prediction models for slopes on the Loess Plateau.