Research on psychostimulants, analysis of animal models and genetic association studies all suggest that the brain-derived neurotrophic factor gene (BDNF) may be a good candidate for pharmacogenetic studies of attention deficit hyperactivity disorder (ADHD). Yet to date there have been no pharmacogenetic studies of BDNF in ADHD. A total of 102 drug-naive ADHD children (8.7±2.1 yr) were treated with osmotic release oral system-methylphenidate (OROS-MPH) for 12 wk, and four kinds of response criteria were applied, based first, on a combined threshold of the ADHD Rating Scale – IV (ARS) and the Clinical Global Impression – Improvement scale (CGI-I); second, on scores of 1 or 2 vs. 3–7 on the CGI – Severity scale; third, on a >50% reduction in ARS scores; and fourth, on satisfaction of all of the aforementioned criteria. The Val66Met polymorphism of BDNF and six single nucleotide polymorphisms from the SLC6A2, ADRA2A and NTF-3 genes were tested for association with each criterion. Relative to other genotypes, homozygosity for the Val allele of the BDNF Val66Met polymorphism was associated with a greater relative frequency of good response under all four response criteria (after controlling for baseline ARS score, age, gender, final dose (mg/kg) of OROS-MPH at 12 wk, and level of academic functioning). This association was significant at the uncorrected level for the first and third response criteria (p=0.013 and p=0.018, respectively) and significant at a Bonferroni-corrected level for the second and fourth response criteria (p=0.0002, p=0.0003, respectively). Our findings support an association between homozygosity for the Val allele of BDNF and better response to OROS-MPH in Korean ADHD children as assessed by four different response criteria.