To verify whether it is possible, in people with diabetes mellitus (DM) considered at very high cardiovascular (CV) risk, stratify this risk better and identify significant modifiable risk factor (including lifestyle habits) to help patients and clinicians improve CV prevention. People with DM and microvascular diseases or one or more CV risk factors (hypertension, hyperlipidemia, smoking, poor dietary habits, overweight, physical inactivity) included in the Risk and Prevention study were selected. We considered the combined endpoint of non-fatal acute myocardial infarction and stroke and CV death. A multivariate Cox proportional analysis was carried out to identify relevant predictors. We also used the RECPAM method to identify subgroups of patients at higher risk. In our study, the rate of major CV events was lower than expected (5% in 5years). Predictors of CV events were age, male, sex, heart failure, previous atherosclerotic disease, atrial fibrillation, insulin treatment, high HbA1c, heart rate and other CV diseases while being physically active was protective. RECPAM analysis indicated that history of atherosclerotic diseases and a low BMI defined worse prognosis (HR 4.51 95% CI 3.04-6.69). Among subjects with no previous atherosclerotic disease, men with HbA1c more than 8% were at higher CV risk (HR 2.77; 95% CI 1.86-4.14) with respect to women. In this population, the rate of major CV events was lower than expected. This prediction model could help clinicians identify people with DM at higher CV risk and support them in achieving goals of physical activity and HbA1c.