The prevalence of diabetes is increasing globally. In addition to established risk factors for diabetes, such as diet, inactivity, overweight and obesity, the involvement of persistent organic pollutants, including dioxins and polychlorinated biphenyls (PCBs), has also been suggested to be a possible, but controversial, cause of this epidemic. The present study investigated the association between blood PCB congener levels and the prevalence of diabetes among middle-aged, overweight and obese Japanese participants in the Saku Control Obesity Program. One hundred seventeen participants had their congener-specific PCB levels measured in addition to undergoing routine blood analyses at the time of a medical checkup. Prevalent diabetes was defined according to two methods: definite diabetes was defined as people with an HbA1c level ≥ 6.9% or who were taking medication for diabetes, and all diabetes was defined as people with an HbA1c level ≥ 6.5%, a fasting plasma glucose level ≥ 126 mg/dL, or a history of doctor-diagnosed diabetes. A multiple logistic regression analysis was performed to analyze the association between the PCB levels and the prevalence of diabetes, with adjustments for sex, age, body mass index and total lipids. As a result, PCB 146 and 180 were positively associated and PCB 163/164 and 170 were negatively associated with the prevalence of definite diabetes. The significance of the association of PCB 180 and 163/164 with the prevalence of diabetes persisted regardless of the definition of diabetes or adjustments for total lipids, suggesting the possibility that these parameters may modify the risk of diabetes.