Metformin is a biguanide antihyperglycemic agent often used for the treatment of non-insulin dependent diabetics (NIDDM). In this study, the pharmacokinetics and pharmacodynamics of metformin were investigated in Korean healthy volunteers during a fasting state for over 10 h. In order to evaluate the amount of glucose-lowering effect of metformin, the plasma concentrations of glucose were measured for a period of 10 h followed by the administration of metformin (oral 500 mg) or placebo. In addition, the concentration of metformin in blood samples was determined by HPLC assay for the drug. All volunteers were consumed with 12 g of white sugar 10 minutes after drug intake to maintain initial plasma glucose concentration. The time courses of the plasma concentration of metformin and the glucose-lowering effect were analyzed by nonlinear regression analysis. The estimated $C_{max}$, $T_{max}$, $CL_{t}$/F (apparent clearance), V/F(apparent volume of distribution), and half-life of metformin were 1.42${pm}$0.07 $mu extrm{g}$/mL, 2.59${pm}$0.18h, 66.12${pm}$4.6 L/h, 26.63 L, and 1.54 h respectively. Since a significant counterclock-wise hysteresis was found for the metformin concentration in the plasma-effect relationship, indirect response model was used to evaluate pharmacodynamic parameters for metformin. The mean concentration at half-maximum inhibition $IC_{50}$, $k_{in}$, $k_{out}$ were 2.26 $mu extrm{g}$/mL, 83.26 $H^{-1}$, and 0.68 $H^{-1}$, respectively. Therefore, the pharmacokinetic-pharmacodynamic model may be useful in the description for the relationship between plasma concentration of metformin and its glucose-lowering effect.