Abstract Background There is lack of information of the hospital costs related to depression. Here, we compare the costs associated with general hospital admissions over 2 years between older men with and without a documented past history of depression. Methods A community-based cohort of older men living in Perth, Western Australia, was assessed at baseline between 2001 and 2004 and followed up for 2 years by prospective data linkage. The participants were selected randomly from the Australia electoral roll. Two-year hospital costs were estimated. Results Among 5411 patients, 75% of 339 men with depressive symptoms had at least one hospital admission compared with 61% of 5072 men without depression (P<.001). Two-year median hospital costs in the depressed group were A$4153 compared with A$1671 in participants free from depression (P<.001). In multivariate analysis, the presence of clinically significant depressive symptoms remained an independent predictor of higher cost [incident rate ratios (RR)=1.44, 95% confidence interval (CI): 1.23–1.68] and was associated with being a high-cost user of health services (RR=2.04, 95% CI: 1.43–2.92). Limitations The estimation of costs was solely based on the main diagnosis, potentially leading to underestimates of the real cost differences. Conclusions Hospital care cost was higher for older men with documented evidence of past depression than those without. The issue of depression in later life must be tackled if we want to optimize the use of limited hospital resources available.