Quantitative Ion Character-Activity Relationship (QICAR) was used for correlating metal ionic properties with maximum biosorption capacity (q(max)). Heat inactivated biomass of Staphylococcus saprophyticus BMSZ711 was studied for biosorption of nine metal ions. Influence of contact time and initial pH was checked. q(max) was determined by Langmuir isotherm and followed a descending sequence (in mmol/g): Pb(2+)>Cd(2+)>Cr(3+)>Zn(2+)>Hg(2+)>Cu(2+)>Co(2+)>Ni(2+)>K(+). q(max) values was modeled with 20 metal ionic characteristics, among these covalent index (X(m)(2)r) was best fitted with q(max) for all metal ions tested, in the following model: q(max)=0.09+0.11(X(m)(2)r) (R(2)=0.73, AIC=-4.14). Classification of metal ions according to valence or soft/hard improved QICARs modeling and more characteristics significantly correlated with q(max) which revealed that covalent bonding played major role in biosorption of soft metal ions and ionic bonding for borderline and hard ions. Biosorption capacity was most effectively predicted (R(2)=0.99, AIC=-8.04) with a two variable model containing electro-negativity (X(m)) and softness index (o(rho)(')).