• • Adsorption capacity order Pb > Ni > Cd was consistent across pH conditions, with maximum adsorption differences up to 0.5 mmol/g at pH 7, driven by hydrated radius and electronegativity.
• • The 1-site/2-pKa SCM fitted adsorption edge data with correlation coefficients R ≥ 0.94, enabling accurate prediction of metal adsorption across a range of soil pH (4–8).
• • Linear regression models linked SCM parameters to soil properties, with pH as the most significant predictor for lgKSOMe (R² > 0.85) and Ds, while free iron oxide content negatively correlated with pKa2 (r = -0.72).
• • Validation on independent soil samples achieved R² between 0.75 and 0.82 and RMSE between 0.1 and 0.51 mmol/g, demonstrating the model's practical utility for rapid risk assessment without exhaustive adsorption experiments.