Sulfate-reducing reactive walls installed in situ in the path of acid mine drainage contaminated groundwater, present a promising passive treatment technology. However, a rigorous and methodical selection of the most appropriate reactive mixture composition still needs to be investigated. The aim of this study was the selection of the most reactive medium using a multiple factor design and the modeling of the sulfate-reduction rate. Reactivity of 17 mixtures was assessed in batch reactors (in duplicates) using a synthetic AMD. Results indicate that within 41 days, sulfate concentrations decreased from initial concentrations of 2,000-3,200 mg/l to final concentrations of <90 mg/l. Metal removal efficiencies ranged between 51-84% for Ni and 73-93% for Zn. Generated sulfate-reduction rate predictive models which had very satisfactory parameters (R2 = 0.86, F = 62.38 (p-level < 10(-13)) and R2 = 0.90. F = 62.30 (p-level < 10(-13))) identified poultry manure and two other carbon sources as the critical variables for sulfate-reduction rate.