The few cohort studies that have investigated the association between exposure to air pollution and occurrence of diabetes have reported conflicting results. We aimed to evaluate the association of long-term exposure to particulate matter (PM), nitrogen oxides (NO) and ozone (O), with baseline prevalence and incidence of type 2 diabetes in a large administrative cohort in Rome, Italy. A total of 1,425,580 subjects aged 35+years (January 1st, 2008) were assessed and followed for six years. We estimated PM, PM, PM, NO, and NO exposures at residence using land use regression models, and summer O exposure using dispersion modeling. To estimate the association between air pollutant exposures and prevalence and incidence of diabetes, we used logistic and Cox regression models, considering individual, environmental (noise and green areas), and contextual characteristics. We identified 106,387 prevalent cases at baseline and 65,955 incident cases during the follow-up period. We found positive associations between nitrogen oxides exposures and prevalence of diabetes with odds ratios (ORs) up to 1.010 (95% CI: 1.002, 1.017) and 1.015 (1.009, 1.021) for NO and NOx, respectively, per fixed increases (per 10μg/m and 20μg/m, respectively). We also found some evidence of an association between NOx and O and incidence of diabetes, with hazard ratios (HRs) of 1.011 (95%CI: 1.003-1.019) and 1.015 (1.002-1.027) per 20 and 10μg/m increases, respectively. The association with O with incident diabetes was stronger in women than in men and among those aged <50years. In sum, long-term exposure to nitrogen oxides was associated with prevalent diabetes while NOx and O exposures were associated with incident diabetes.