Anesthetic doses of fentanyl are sometimes determined empirically by repetetive doses which are accompanied by fluctuating concentrations. If a constant plasma concentration can be maintained the analgesis and anesthesia can be guaranteed for the duration of surgery. After an I.V. bolus injection (3mcg/kg), fentanyl (2mcg/kg/kr) was continuously infused for an average length of time of 78.3minutes in a manner analogous to the safety of the administration of inhalation anesthetics. To evaluate whether this method is applicable to the clinical situation, the decay of plasma fentanyl with time was simulated based on a population pharmacokinetic model used to predict estimated concentrations in the plasma. The clinical effect of the concentrations was determined by a scoring system which consisted of 4 components, systolic blood pressure, heart rate, sweat production and lacrimation. Using a modified computer program, the mean and 68% confidence interval of the estimated fentanyl plasma concentration (peak concentration 8.5~6.8 mcg/I, maintenance concentration 1.3~0.8 mcg/I, 10 minutes after discontinuation of infusion 0.9~0.5 mcg/I and elimination half-life 267±15.7 minutes) were predicted, while the clinical scores were maintained under 4 points. The pharmacokinetic model based on estimation proved to be effective and valid for all patients and predicted a steady target concentration of fentanyl in the plasma.