Purpose The use of exercise as an intervention to improve health in the general population is well documented. The purpose of this paper is to explore whether an exercise referral scheme can be an effective health promotion tool for male prisoners in Ireland, presenting with mental health symptoms. Design/methodology/approach This mixed methods study with a pre- and post-intervention design was conducted in Mountjoy Prison, Dublin, which has a capacity for approximately 790 prisoners. Reliable and validated symptom assessment scales were used to assess levels of depression, anxiety, stress, self-esteem and anger amongst a sample of 40 prisoners pre- and post-intervention. The scales used were the Depression, Anxiety and Stress scale or DASS-42 (Lovibond and Lovibond, 1995), the Novaco Anger Scale (Novaco, 1994), the Rosenberg Self-Esteem Scale ( Rosenberg, 1965 ) and the Zung Self-Rated Anxiety Scale (Zung, 1971). Semi-structured interviews were also conducted with a subset of the participants post-intervention to further test and contextualise the symptom ratings. The data gathered from the self-rating scales were imported into SPSS 22 for statistical testing for significance. Wilcoxon's signed-rank test was then used to measure significance of changes. Thematic analysis was performed on the qualitative data. Findings In the post-intervention, significant levels of improvement were achieved in the levels of depression, anxiety (DASS), anxiety (Zung), stress, anger, and self-esteem for 29 of the 30 prisoners who completed the study. The incidence of normal mood scores rose from 33 to 90 per cent after the intervention; the incidence of extremely severe scores for anxiety changed from 40 to 7 per cent, severe stress scores changed from 27 to 3 per cent, normal stress levels rose from 17 to 73 per cent, marked anger ratings reduced from 40 to 3 per cent and low self-esteem levels reduced from 20 per cent of participants pre-intervention to 7 per cent post-intervention. In the main, participants perceived the experiences and outcomes of the intervention positively. Research limitations/implications There are some limitations to the design of this study. Operational circumstances within the prison at the start of this study prevented the authors from accessing a larger sample. A control group would add greatly to the study but this was not possible within a single prison setting. The possible influence of extraneous variables such as increased attention and social contact, and more time out of one's cell may have contributed to improved symptom scores as much as the exercise intervention in this study. This possibility was recognised from the outset but the authors proceeded because the aim was to test if an exercise referral package (and all that inevitably goes with that) would make a difference for symptomatic prisoners. Practical implications The organisation and smooth running of the intervention and the positive results therein underpinned the practicality of this project. The significantly positive results con...