OBJECTIVE To study the correlation between serum and salivary phenytoin concentration in Thai epileptic children. MATERIAL AND METHOD Children aged 5 to 12 years with diagnosed epilepsy who received phenytoin monotherapy seen in the neurological clinic at Queen Sirikit National Institute of Child Health were studied. The recruited patients were required to have good compliance, normal albumin level, and no evidence of cancer, HIV infection, hepatic, renal and salivary glands disease. Blood and saliva samples were collected and measured phenytoin level by fluorescence-polarization immunoassay technique. RESULTS Thirty patients, 19 males and 11 females, were studied. The average (mean +/- SD) age and weight were 8.24 +/- 2.09 years and 27.76 +/- 9.86 Kilograms. Both serum and salivary phenytoin levels correlated with phenytoin doses as exponential type (R2 = 0.4188, 0.3682, respectively). Equations for describing serum and salivary phenytoin levels by phenytoin dose were y = 0.7403e(0.3952x) and y = 0.1431e(0.3072x) respectively. Serum and salivary phenytoin levels were closely correlated as linear type (R = 0.880, R2 = 0.967). The obtained equation of this relationship was y = 10.165x, where y = serum phenytoin level and x = salivary phenytoin level. Adverse drug reactions were found in 5 patients (6.6%), horizontal nystagmus 2 cases, hirsutism 2 cases and gingival hyperplasia 1 case. CONCLUSION High correlation between serum and salivary phenytoin levels supported the use of saliva instead of blood for phenytoin monitoring in Thai children which were difficult in blood collection and had psychological trauma. The obtained equations in the present study could be applied for adjusting the dosage regimen and monitoring by using salivary phenytoin level in clinical practice.