The aim of this paper is three-fold. First, we identified self-and informant-rated Five-Factor Model (FFM) personality domains and facets associated with diabetes diagnosis. Second, we tested whether the associations were independent of the rater method-specific variance.Lastly, we examined whether the observed associations were mediated by BMI, alcohol intake, dietary habits, and exercise. The participants were members of the Estonian Biobank (N = 3592; 1145 men; Mage = 46.6 ± 7.0 years). We fit a series of logistic regression models predicting diabetes diagnosis from one self-or informant-rated personality domain or facet at a time, controlling for age, sex, and education. Diabetes diagnosis was significantly associated with the N5: Impulsiveness, E4: Activity, and C2: Order facets. Methodindependent variance, estimated by means of bi-factor models, was significantly related with diabetes for two of the facets, E4: Activity (β = -0.106, p = .007) and C2: Order (β = -0.089, p = .037), but not for N5: Impulsiveness. The strongest mediator of the personality-diabetes association was BMI, explaining 30-50% of the observed associations. We discuss implications of the current results.