The data obtained from synthetic mixtures of metal ions were processed by radial basis function networks (RBFNs) and back-propagation neural network. The optimal conditions of the neural networks were obtained by adjusting various parameters. Satisfactory precision and accuracy were obtained with all networks, although, because of surprisingly lower root mean square error (%) values, RBFNs were the preferred approach. The proposed approach was tested by analysing the composition of the different mixtures containing Zn(2+) , Mn(2+) and Fe(3+) . The proposed method was successfully applied to the simultaneous determination of Zn(2+) , Mn(2+) and Fe(3+) ions in milk and vegetable samples.