Attention-deficit hyperactivity disorder (ADHD) is a common behavioral disorder observed during childhood, detected in 3% to 5% of school-age children. The disorder is characterised by marked inattention, hyperactivity, and impulsiveness. In most cases, symptoms can be treated by catecholamine-releasing drugs, such as methylphenidate. Children with ADHD are at higher risk for substance abuse and oppositional, conduct and mood disorders. Familial and adoption studies shed light on the genetic vulnerability of ADHD. Twin studies estimated the broad heritability to range between 40% and 90%. The mode of transmission is yet unknown, but is likely polygenic. Molecular genetic studies in ADHD should contribute to a greater understanding of the pathophysiology of the disorder (genetics of the vulnerability), and could help to select a more rational type of treatment (pharmacogenetic). Family-based association studies already performed are reviewed in this manuscript. Association studies, using haplotype relative risk (HRR) or transmission disequilibrium test (TDT) have focused on candidate genes which code for proteins potentially involved in the etiopathogenesis of the disorder. Genes involved in dopamine, serotonin, and noradrenalin systems have thus been assessed for their role in core features of ADHD, such as motor overactivity, inattention, and impulsiveness. According to a meta-analysis, the DAT1 gene, an obvious candidate gene in ADHD vulnerability, does not appear to be involved (OR = 1.13, p = 0.21). On the other hand, DRD4 (OR = 1.26, p = 0.01) and DRD5 (OR = 1.4, p = 0.01) are significantly associated to ADHD according to the present meta-analysis, confirming previous ones. Recent studies showed a trend for an association between one allele of the 5-HTT (considering case-control studies) and DBH (OR = 1.27, p = 0.06) genes and ADHD, but these positive findings have to be replicated. ADHD is a complex disorder with potentially many different risk factors. Genetic and phenotypic heterogeneity could explain why some association studies are positive, whereas others are negative. For instance, different developmental pathways are likely to lead to similar clinical outcomes. More clear-cut phenotypes, such as ADHD with conduct disorder, or ADHD with bipolar disorder, could be more homogenous, the genes involved being therefore more easy to detect. These phenotypes are beginning to be specifically studied in molecular genetics. In addition, the development of pharmacogenetics could help to identify predictors of clinical response for a specific type of treatment, which would be clearly helpful in clinical practice.