Background Studies of the overlap between severe mental disorder and criminality tend to focus on prison populations rather than psychiatric populations. Aims Our aims were to establish the prevalence of previous imprisonment among female psychiatric inpatients and test relationships between diagnoses, mortality and imprisonment. Methods A nationwide cohort of 18–65‐year‐old women who had been hospitalised for psychiatric disorder between January 1983 and March 2008 was identified from a hospital records database and linked to the database of the Prison and Probation Administration of Iceland as well as the National Register of causes of death at Statistics Iceland from January 1985. Results Six thousand and ninety‐four women had had at least one psychiatric hospitalisation, 102 of them had been imprisoned on 172 occasions between them, giving an imprisonment rate of 118 per 100,000 over the 24 year period of study. The crude imprisonment proportion was 1.7% during a 20‐year follow‐up period; it was at its peak (5%) among 18–30 year‐olds at index admission. Substance use and personality disorders were the most common diagnoses associated with imprisonment. Mortality rates were not statistically different between those imprisoned and not (hazard ratio = 1.3, 95% confidence interval 0.5–3.5). Conclusion and implications for practice Women admitted to a psychiatric hospital have higher rates of imprisonment than the general population. Because admission predated imprisonment in most cases, this may be seen as an opportunity for early intervention to reduce later criminality. Copyright © 2015 John Wiley & Sons, Ltd.