The objective of the present work was to evaluate the practicability of several procedures used for prediction of the serum concentration of lithium in a steady state which will make it possible to adjust the dosage in the early period of treatment. The group comprised nine men where prophylactic lithium treatment was indicated on account of maniomelancholy. The authors assessed in each patient the theoretically calculated and actually assessed value of minimal lithium concentrations in a steady state. In all applied predictive methods the predicted values were always higher than the actually assessed lithium concentrations. The least suitable method was that based on individualization of the population average of the velocity constant of elimination. The predictive value of the procedure based on a non-individualized single-point estimate was higher. A statistically significant correlation of predicted and assessed values was provided by the method which, based on two collected samples of lithium serum concentrations, makes it possible to estimate the individual value of the cumulative factor.