Objectives: FRAX is a computer-based algorithm developed by the World Health Organisation for estimation of the 10-yr risk of a hip or major osteoporotic fracture. Inclusion of femoral neck bone mineral density (BMD) in the estimation is optional. The study aimed to investigate the intra-individual agreement between FRAX fracture risk calculated with and without BMD in patients with rheumatoid arthritis (RA). Methods: Clinical data and BMD results from 50 RA patients registered in the Danish rheumatology registry (DANBIO) were used for analysis. Using the Bland–Altman method, lower and upper 95% limits of agreement [LLoA;ULoA] between intraindividual assessments of fracture risk with and without BMD and the bias (mean of individual differences) were calculated. Categorization of patients according to the National Osteoporosis foundation (NOF) treatment thresholds were also assessed with and without BMD. Results: Mean age was 63.6 ± 11.7 yr, mean disease activity score (DAS28-CRP) 3.3 ± 3.5 and mean femoral neck T-score −1.43 ± 1.15. The mean 10-yr risk of a major fracture and a hip fracture calculated with BMD was 22.9 ± 15.8% and 8.5 ± 10.8%, respectively. The LLoA and ULoA [bias] calculated without vs with BMD were −14.5 and 20.4 percent point (pp) [2.9 pp] for major fracture risk and −14.0 and 23.2 pp [4.6 pp] for hip fracture. NOF treatment categorization was only dependent on BMD in 4% of the patients. Conclusion: The FRAX fracture risk estimated with and without BMD may disagree substantially in individual patients with RA but this seems to have only little impact on treatment categorization based on the NOF guidelines.