BACKGROUND AND AIMS: Diabetes is a risk factor for cardiovascular disease (CVD), yet southern European migrants to Australia with high rates of type 2 diabetes have relatively low CVD mortality. Our aim was to determine whether a Mediterranean style diet could reduce mortality in people with diabetes. METHODS AND RESULTS: Participants included 16,610 males and 23,860 females from the Melbourne Collaborative Cohort Study; 25% were born in Greece or Italy, and 2150 had previously been diagnosed with diabetes or had elevated blood glucose at baseline (1990-94). Data on demographic, behavioral and physical risk factors were also collected. A personal Mediterranean Diet Score (MDS) was calculated using data from a validated 121-item food frequency questionnaire. Total and CVD mortality data were available up to 2003. Diabetes (new and known) at baseline, was associated with total mortality (men HR 1.43, 95%CI 1.26-1.62; women HR 1.86 95%CI 1.58-2.18), and CVD mortality (men HR 1.53, 95%CI 1.21-1.94; women HR 2.10 95%CI 1.48-2.97) in multivariate models. There was no evidence that glucose tolerance modified the associations between MDS and total or CVD mortality (p interaction all > 0.16). The HRs for total mortality per unit of MDS were 0.96 (95% CI 0.93-0.99) in men and 0.94 (95% CI 0.92-0.97) in women. The HRs for CVD mortality per unit of MDS were 0.94 (95% CI 0.89-0.99) in men and 0.94 (95% CI 0.87-1.01) in women. CONCLUSION: Our results add to the evidence supporting the benefit of a Mediterranean style diet for people with type 2 diabetes.