The influence of metal ionic characteristics on their biosorption capacity was analyzed using quantitative structure-activity relationships model. The waste biomass of Saccharomyces cerevisiae was used as biosorbent to adsorb 10 kinds of metal ions, and their maximum biosorption capacity (qmax) was determined by the Langmuir isotherm model. The values of qmax decreased in the following order (in millimole per gram): Pb2+ (0.413) > Ag+ (0.385) > Cr3+ (0.247) > Cu2+ (0.161) > Zn2+ (0.148) > Cd2+ (0.137) > Co2+ (0.128) > Sr2+ (0.114) > Ni2+ (0.108) > Cs+ (0.092). Twenty-two parameters of physiochemical characteristics of metal ions were selected and correlated with qmax, i.e., OX, AN, r (Å), ΔIP (eV), ΔE0 (V), Xm, |log KOH|, Symbol, Z2/r, AN/ΔIP, σρ, AR, AW, IP, AR/AW, Z/r2, Z/AR2, Z/r, Z/AR, Z*2/r·, Z*, N. The linear regression analysis showed that the covalent index Symbol was correlated well with qmax for all metal ions tested in the following equation: qmax = 0.029 + 0.061 (Symbol) (R2 = 0.70). It suggested that the greater the covalent index value of metal ion was, the greater the potential to form covalent bonds with biological ligands, such as sulphydryl, amino, carboxyl, hydroxyl groups, etc. on the biomass surface, and the higher the metal ion biosorption capacity was. Classification of metal ions, for divalent ion or for soft-hard ion could improve the linear relationship (R2 = 0.89). The equation could be used to predict the biosorption capacity of metal ions.