Background Smokers, previous smokers and a never smoked group were compared on self-reported and real world prospective memory (PM – the cognitive ability of remembering to carry out particular actions at some future point in time). Methods Twenty-seven current smokers, 24 people who had never smoked and 18 previous smokers were compared using an existing groups design. Scores on the long and short term PM subscales of the Prospective and Retrospective Memory Questionnaire (PRMQ) and scores on a Real World Prospective Memory Task (RWPMT) constituted the dependent measures. Smoking and other drug use were assessed by a Recreational Drug Use Questionnaire. The Hospital Anxiety and Depression Scale gauged levels of anxiety and depression. The National Adult Reading Test measured IQ, and retrospective memory was measured using the PRMQ. Gender, age, anxiety and depression, IQ, alcohol use and the retrospective memory scores, were measured as covariates and controlled for in the analysis. Results A series of univariate ANCOVAs were applied to the main PM data across the three groups, controlling for variations in age, gender, mood, IQ, alcohol use and retrospective memory scores. These revealed no significant between-group differences on self-reported PM; however smokers recalled significantly fewer action–location combinations than the never smoked and previous smoker groups on the objective RWPMT. Conclusions Existing smokers showed reduced performance on RWPMT when compared to the never smoked group and previous smokers. Real-world PM impairments should be added to a growing list of neuropsychological sequelae associated with persistent smoking.