Aims. To compare the cognitive performance of methadone maintenance patients (MMPs) and a matched sample of non-heroin-using control subjects, and to ascertain risk factors for poorer cognitive performance. Design. Matched control study. Setting. Sydney, Australia. Participants. Thirty MMPs and 30 non-heroin using controls, matched for age, gender and education. Findings. The MMP group had significantly higher rates than controls of alcohol dependence, heroin overdose and head injury. There was no difference between the groups on pre-morbid functioning. The MMP group performed significantly poorer than controls on all of the neuropsychological domains measured: information processing, attention, short-term visual memory, delayed visual memory, short-term verbal memory, long-term verbal memory and problem solving. A life-time diagnosis of alcohol dependence and the number of non-fatal heroin overdoses were independent significant predictors of poorer cognitive performance. Conclusions. In addition to high rates of psychiatric morbidity, MMPs also show cognitive deficits compared to matched controls. A history of alcohol dependence and repeated exposure to overdose increase the likelihood of cognitive impairment. The current study does not rule out the possibility of other factors, that were not measured, that may contribute to cognitive impairment among this MMPs.