Background & Aims: The effects of sex and menopausal status on the association between NAFLD and incident type 2 diabetes (T2D) remain unclear. We investigated the effect modification by sex and menopause in the association between NAFLD and T2D; also, added predictive ability of NAFLD for the risk of T2D was assessed. Approach & Results: This cohort study comprised 245,054 adults without diabetes (109,810 premenopausal women; 4,958 postmenopausal women; 130,286 men). Cox proportional hazard models were used to estimate hazard ratios (HRs; 95% confidence intervals [CIs]) for incident T2D according to NAFLD status. The incremental predictive role of NAFLD for incident T2D was assessed using the area under the receiver operating characteristic curve, net reclassification improvement, and integrated discrimination improvement. A total of 8,381 participants developed T2D (crude incidence rate/103 person-years: 2.9 premenopausal women; 12.2 postmenopausal women; 9.3 men) during median follow-up of 5.3 years. NAFLD was positively associated with incident T2D in all groups. After adjustment for potential confounders, the multivariable-adjusted HRs (95% CIs) for incident T2D comparing NAFLD to no NAFLD were 4.63 (4.17–5.14), 2.65 (2.02–3.48), and 2.16 (2.04–2.29) in premenopausal women, postmenopausal women and men, respectively. The risks of T2D increased with NAFLD severity as assessed by serum fibrosis markers, and the highest relative excess risks were observed in premenopausal women. The addition of NAFLD to conventional risk factors improved risk prediction for incident T2D in both sexes, with a greater improvement in women than men. Conclusions: NAFLD, including more severe NAFLD, is a stronger risk factor for incident T2D in premenopausal women than in post-menopausal women or men; protection against T2D is lost in pre-menopausal women with NAFLD.