In this paper a methodological approach is proposed to validate mechanistic modeling for proton binding onto active sites of mineral and soil samples by reducing the uncertainty and arbitrariness of model schematization. This approach is based on the quantitative formulation (X-ray calibration method) of a simulating mineral mixture (SMM) accounting for the main mineral phases in the soil (quartz, goethite, hematite, muscovite, clinochlore). Mineral and organic contributions were separated by comparing titration curves of river sediment and SMM. Specific mineral contributions to the acid properties of SMM were separated by comparing titration models of SMM and single minerals. Different nonelectrostatic models were used for titrations of SMM and single minerals: two-site/three-KH models (one amphoteric plus one monoprotic site) for clay minerals and SMM; one-site/two-KH models (one amphoteric site) for goethite and hematite; and a one-site/one-KH model (one monoprotic site) for quartz. Crossed-comparisons of titration models allow for identifying and quantifying the specific contributions of the distinct edge hydroxyl groups of iron oxides, clay minerals, and quartz in the different pH ranges. In particularthe amphoteric sites of aluminosilicates mainly contribute in the acid-neutral pH range, the amphoteric sites of iron oxides take part in the neutral-basic range, and finally the monoprotic edge hydroxyl groups of quartz react in the upper basic region of pH. The good simulation of the acid-base properties of SMM (according to single mineral titration models and quantitative composition by X-ray) confirms both model schematization and SMM formulation. Speciation diagrams of the active sites of the different mineral components (aluminosilicates, iron oxides, and quartz) were obtained by implementing the database of a dedicated software with the apparent equilibrium constants regressed by titration modeling of single minerals.