The purpose of this population-based prospective cohort study was to examine the effect of hormone therapy (HT) on incidence of diabetes mellitus (DM). Eight thousand four hundred and eighty-three DM-free post-menopausal women aged 52-62 from the population-based Kuopio osteoporosis risk factor and prevention study were followed for 5 years from 1994-1999. Information about the use of HT and health events was obtained from three repeated questionnaires in 1989, 1994, and 1999. DM morbidity before and during the follow-up was obtained from the Registry of Specially Refunded Drugs of the Finnish Social Insurance Institution. Kaplan-Meyer survival curves and Cox's proportional-hazards models were used to estimate the risk of incident DM in relation to the use of HT. During the follow-up, 40.8% DM-free post-menopausal women had never used HT, 27.3% women were HT past users and 31.9% women had used HT presently during the follow-up. During the follow-up, 162 incident DM cases were recorded. Compared with never users of HT, the adjusted hazard ratio of DM was 0.81 (95% confidence interval (CI) 0.57-1.16) for only past users, 0.53 (95% CI 0.24-1.15) in part-time (during the follow-up <2.5 years) users and 0.31 (95% CI 0.16-0.60) in continuous (during the follow-up 2.5-5.0 years) users of HT. HT use decreases the incidence of DM in post-menopausal women.