Continuous exposure to low doses of persistent organic pollutant (POPs), such as those occurring in the general population, might contribute to the burden of type 2 diabetes mellitus (T2DM). However, evidences from longitudinal studies are scarce. We aimed to explore the associations of accumulated POP exposure with the development of T2DM by means of 1) longitudinal associations with the 16-year incidence of the disease, and 2) complementary cross-sectional analyses with markers of glucose homeostasis at recruitment. Organochlorine pesticide and polychlorinated biphenyl (PCB) concentrations were analyzed in adipose tissue samples and incident T2DM cases were retrieved from clinical records. Homeostatic model assessment values of insulin sensitivity/resistance and β-cell function at recruitment were calculated. Linear and Cox-regression models were performed. In individuals with normal weight/overweight (n = 293), we observed positive dose–response relationships between the studied POPs and T2DM risk, particularly for hexachlorobenzene (HCB) [hazard ratio (HR): 3.96 for 4th quartile versus 1st quartile (Q1); confidence interval (CI) 95%: 0.79, 19.71]. PCB-180 showed a positive but seemingly non-linear association with T2DM risk [HR of 3er quartile (Q3) versus Q1: 6.48; CI 95%: 0.82, 51.29]. Unadjustment for body mass index considerably increased the magnitude of the associations. In the cross-sectional study (n = 180), HCB and PCB-180 were inversely associated with insulin sensitivity and positively associated with insulin resistance parameters. Our results suggest that a higher burden of specific POPs in adipose tissue may disrupt glucose homeostasis, possibly contributing to increase T2DM risk, especially in non-obese adults.