This study aimed to compare the performance of Fracture Risk Assessment Tool (FRAX) with that of Derived FRAX (DeFRA) in estimating fracture risk in a cohort of type-2 diabetes mellitus (T2DM) postmenopausal women. One hundred nineteen T2DM postmenopausal women and 118 consecutive healthy postmenopausal women were enrolled. Fracture risk was assessed with FRAX (adjusted or non- for trabecular bone score, TBS) and DeFRA. Bone mineral density (BMD) and TBS were evaluated by dual-energy X-ray absorptiometry (DXA). The outcome was the presence of vertebral/non-vertebral fragility fractures (FFs). T2DM women showed higher spinal BMD T-score (p < .05), but lower TBS (p < .05), than controls. Diabetic patients had higher prevalence of FFs compared to controls (p < .05), but no significant difference were found in the scores of any of the predictor tools. Differently, in the T2DM group, the scores of DeFRA, FRAX and adjusted-FRAX were significantly (p < .01 for all) higher in fractured compared with non-fractured women. DeFRA showed the best discriminative power among all fracture risk predictor tools (area under curves: DeFra: 0.89; adjusted FRAX: 0.80; non-adjusted FRAX: 0.73). In summary, all fracture risk assessment tools appeared to be effective in predicting bone fractures in T2DM postmenopausal women, with DeFRA showing a slightly better diagnostic accuracy.