Routine clinical pharmacokinetic data collected from outpatients receiving phenytoin (PHT) were reanalysed to estimate population pharmacokinetic parameters. There were 756 steady-state PHT concentrations and associated dosage rates (mg/d) from 334 outpatients. The data were analysed using nonlinear effects model (NONMEM), a computer program designed for population pharmacokinetic analysis that allows pooling of data from many individuals. The influence of weight, co-anticonvulsants on the maximum elimination rate (V_m) and age, co-anticonvulsants on the Michaelis-Menten constant (K_m) and the influence of dosage form on the bioavailability (F) of PHT were investigated.The V_m and K_m of a 60 kg adult outpatient treated with PHT alone were estimated to be 325 mg/d and 2.41 μg/ml, respectively, while for a same size individual taking PHT with co-anticonvulsants respective estimates were 351 mg/d and 3.18 μg/ml. The parameter of a power function of weight was estimated to adjust V_m for body size. The best function adjusts V_m in proportion to weight to the 0.737 power. The K_m for patients less than 15 years old was 24.8% less than that of adults. Assuming the F of PHT to be 1 in patients prescribed a tablet, the F value in patients prescribed a powder was [1-exp(-9.92/D_ij)]; D_ij is the daily dose of PHT for the ith C_pss in the jth patients (mg/kg/d).