In this study, a multi-component sorption equilibria calculation, with application of artificial neural network (ANN) and identification of adsorption dynamics model using evolutionary algorithm (EA), is presented. Equilibrium experiments were carried out to estimate sorptivity of a new form of a chitosan foamed structure and its selectivity towards Cu(II). Zn(II) and Cr(VI) ions. In the case of single ions, it was found that in the whole range of concentrations, experimental data were well described by the Langmuir-Freundlich equation. In the case of a multi-component mixture the application of a neural MLP network was proposed. Calculations with the use of MLP enabled description of sorption isotherms for when one, two and three ions were present at the same time in the solution. The network also enabled an analysis of sorption of the selected ion, taking into account the effect of its concentration on the sorption of other ions. This assessment would not be possible in an experimental way only. A universal mathematical model of adsorption in a packed column is proposed in this paper. The model includes mass balances for fluid and adsorbent as well as a sorption kinetics. The effect of these, is a system of two partial differential equations. Additionally, the distance and time are composed in one relevantly defined variable. The proposed transformations convert the system of partial differential equations to a system of ordinary equations, which enables analytical solution of the equations system. Also, calculation of a concentration distribution within the solution and adsorbent, dependent on the distance from inlet, and process duration, is achieved. The data obtained in the measurements for Cu(II). Ni(II) and Zn(II) ions, were then compared with those obtained from the model using EA for identification of model coefficients. (c) 2010 Elsevier Ltd. All rights reserved.