Elevated blood pressure (BP) may lead to incident diabetes. However, data about the effect of different BP components on incident diabetes in Middle Eastern women is lacking. We evaluated systolic BP (SBP), diastolic BP (DBP), pulse pressure (PP) and mean arterial pressure (MAP) as independent predictors of diabetes in Iranian women. We performed a population-based prospective study among 3028 non-diabetic women, aged ≥20 years. Odds ratios (ORs) of diabetes were calculated for every 1 s.d. increase in SBP, DBP, PP and MAP. During ≈6 years of follow-up, 220 women developed diabetes. There were significant interactions between family history of diabetes and SBP, PP and MAP (P≤0.01) in predicting incident diabetes. In women without a family history of diabetes, all BP components were significantly associated with diabetes in the age-adjusted model; the risk factor-adjusted ORs were significant (P<0.05) for SBP, PP and MAP (1.30, 1.34 and 1.27, respectively) with similar predictive ability (area under the receiver operating characteristic curve ≈83%). In women with family history of diabetes, in the age-adjusted model, SBP, DBP and MAP were associated with diabetes; in multivariable model, they were not independent predictors of diabetes. In conclusion, in women without family history of diabetes, SBP, PP and MAP, were independent predictors of diabetes with almost similar predictive ability; hence, in the evaluation of the risk of BP components for prediction of diabetes, the presence of family history of diabetes should be considered.