To construct a population pharmacokinetic model for temozolomide (TMZ), a novel imidazo-tetrazine methylating agent and its metabolites MTIC and AIC in infants and children with primary central nervous system tumors. We evaluated the pharmacokinetics of TMZ and MTIC in 39 children (20 boys and 19 girls) with 132 pharmacokinetic studies (109 in the training set and 23 in the validation set). The median age was 7.1 years (range 0.7 to 21.9 years). Children received oral TMZ dosages ranging from 145 to 200 mg/m2 per day for 5 days in each course of therapy. Serial plasma samples were collected after the first and fifth doses of the first and third courses. Approximately eight plasma samples were collected up to 8 h after each dose, and assayed for TMZ, MTIC, and AIC by HPLC with UV detection. A one-compartment model was fitted to the TMZ and metabolite plasma concentrations using maximum likelihood estimation. Covariates, including demographics and biochemical data were tested for their effects on TMZ clearance (CL/F) and MTIC AUC utilizing a two-stage approach via linear mixed-effects modeling. The population mean (inter- and intrapatient variability expressed as %CV) for the pharmacokinetic parameters (based on the training set) were as follows: TMZ CL/F 5.4 l/h (53.4, 17.5), Vc/F 14.0 l (48.5, 39.2), Cmax 9.1 mg/l (20.8, 29.1), and MTIC AUC 1.0 μg/ml·h (13.9, 30.0). Covariate analysis showed that increasing age and body surface area (BSA) were associated with a significant increases in TMZ CL, Vc, and Cmax (P<0.05), and that increasing age was associated with significant decreases in TMZ and MTIC AUC. Indicators of liver and renal function were not significantly associated with TMZ pharmacokinetics or MTIC AUC. The final model with the significant covariates was validated using the remaining 23 pharmacokinetic studies. This study extends previous work done in adults, and identified BSA and age as covariates that account for variability in TMZ disposition in infants and children with primary CNS malignancies.