This study tested whether opiate dependence, tobacco smoking, or their combination accompanied impaired performance on the gambling task (GT), which tests decision-making. GT previously detected impairments in patients with lesions of the ventromedial prefrontal cortex and in substance abusers. Four groups were matched on demographic characteristics and intelligence: methadone-maintained smokers (n = 9) and nonsmokers (n = 9), and control (i.e., not opiate-dependent) smokers (n = 9) and nonsmokers (n = 10). The Wisconsin Card Sorting Task (WCST) was administered to test whether differences in GT performance reflected generalized deficits in prefrontal cortical function. While there were no significant group differences on the WCST, groups differed significantly on GT performance (F(3,31) = 2.95, P = 0.048), controlling for depressive symptom ratings and childhood attention deficit hyperactivity disorder. Methadone-maintained smokers (but not nonsmokers) performed more poorly than either of the two control groups (P = 0.007 versus smokers; P = 0.024 versus nonsmokers). In a planned analysis of methadone-maintained subjects, smokers scored more poorly on GT than nonsmokers (F(1,18) = 5.64, P = 0.032) and had more treatment failures (67% heroin use during the last 30 days versus 20%). The findings suggest that among opiate-dependent individuals, tobacco smoking may be a marker for a more severe form of substance abuse disorder, reflecting impaired decision-making, as modeled by GT.