Considerable evidence indicates that smoking behavior is under a degree of genetic influence. We conducted a systematic review of candidate gene studies of smoking behavior and, where sufficient studies existed, combined reported data using meta-analytic techniques. A total of 41 studies were identified by the search strategy, of which 28 contributed to the meta-analysis. The meta-analysis included data on the DRD2, DAT, 5HTT, and CYP2A6 genes and smoking behavior. Categorical data were extracted on smoking status (never-smoker, ex-smoker, current smoker). Continuous data were extracted on number of cigarettes smoked per day. Evidence indicated effects of the DRD2 Taq1A polymorphism and smoking initiation, the 5HTT LPR and CYP2A6 reduced-activity polymorphisms and smoking cessation, and the DRD2 Taq1A and CYP2A6 reduced-activity polymorphisms and cigarette consumption. The evidence for an effect of specific genes was modest, however, and evidence indicated substantial between-study heterogeneity in most cases, with the exception of the effects of the 5HTT and CYP2A6 genes on smoking cessation. When a random-effects model was applied to analyses in which evidence indicated significant heterogeneity, the effects were in all cases no longer statistically significant. The evidence for a contribution of specific genes to smoking behavior remains modest. Implications for the design of future studies are discussed, such as the need for the development of more specific phenotypes to increase the genetic signal in candidate gene studies.