The aim of this work is to ascertain the potentials of different agroforestry wastes to be used as biosorbents in the removal of a mixture of heavy metals. Fern (FE), rice husk (RI) and oak leaves (OA) presented the best removal percentages for Cu(II) and Ni(II), Mn(II) and Zn(II) and Cr(VI), respectively. The performance of a mixture of these three biosorbents was evaluated, and an improvement of 10% in the overall removal was obtained (19.25 mg/g). The optimum mixture proportions were determined using simplex-centroid mixture design method (FE:OA:RI = 50:13.7:36.3). The adsorption kinetics and isotherms of the optimised mixture were fit by the pseudo-first order kinetic model and Langmuir isotherm. The adsorption mechanism was studied, and the effects of the carboxylic, hydroxyl and phenolic groups on metal-biomass binding were demonstrated. Finally, the recoveries of the metals using biomass were investigated, and cationic metal recoveries of 100% were achieved when acidic solutions were used. (c) 2015 Elsevier Ltd. All rights reserved.