Aims: To explore exposure to 22 persistent organic pollutants (POPs) and incident type 2 diabetes in a population-based, prospective cohort. Methods: This case-cohort study on 753 participants without type 2 diabetes at baseline, was followed-up over nine years, as part of the French D.E.S.I.R. cohort. We measured 22 POPs in fasting serum at baseline. The associations between baseline POP concentrations, pre-adjusted for lipids, BMI, age and sex, with incident type 2 diabetes, were assessed using Prentice-weighted Cox regression models (time scale: age), adjusted for traditional confounding factors. POPs were also modelled summed in functional groups: polychlorinated biphenyls (∑PCB) and organochlorines (∑OC) and also individually, after log-transformation, in adjusted Cox models. Results: There were 200 incident diabetes cases over nine years. Pre-adjusted POP concentrations were not related to diabetes risk for any of the 22 POPs examined. The fully-adjusted hazard ratios (HRs) per interquartile range of the pre-adjusted POPs, ranged from 0.87 (95% CI: 0.64,1.19) to 1.22 (0.93,1.59,). For dichlorodiphenyldichloroethylene (p, p'-DDE) and dichlorodiphenyltrichloroethane (p, p'-DDT), the HRs were 1.09 (0.83,1.43) and 0.89 (0.70,1.13), respectively. The HRs for PeCB, HCB, β-HCH, γ–HCH, oxychlordane, trans-nonachlor were 0.98 (0.85,1.13), 1.06 (0.84,1.33), 1.22 (0.93,1.59), 1.13 (0.89,1.42), 1.00 (0.76,1.31), 0.86 (0.66,1.13), respectively. HRs for ∑PCB, ∑OC and for individual log-transformed POPs did not differ significantly from one. Conclusion: We did not observe any relations between exposure to POPs and diabetes in this population-based cohort. These results do not support causal inferences reported in previous studies linking serum POP concentrations and diabetes risk.