Bone marrow adiposity (BMA) is associated with aging and osteoporosis, but whether BMA can predict bone loss and fractures remains unknown. Using data from the Age Gene/Environment Susceptibility (AGES)‐Reykjavik study, we investigated the associations between ¹H‐MRS–based measures of vertebral bone marrow adipose tissue (BMAT), annualized change in bone density/strength by quantitative computed tomography (QCT) and DXA, and secondarily, with incident clinical fractures and radiographic vertebral fractures among older adults. The associations between BMAT and annualized change in bone density/strength were evaluated using linear regression models, adjusted for age, body mass index (BMI), diabetes, estradiol, and testosterone. Cox proportional hazards models were used to evaluate the associations between baseline BMAT and incident clinical fractures, and logistic regression models for incident vertebral fractures. At baseline, mean ± SD age was 80.9 ± 4.2 and 82.6 ± 4.2 years in women (n = 148) and men (n = 150), respectively. Mean baseline BMAT was 55.4% ± 8.1% in women and 54.1% ± 8.2% in men. Incident clinical fractures occurred in 7.4% of women over 2.8 years and in 6.0% of men over 2.2 years. Incident vertebral fractures occurred in 12% of women over 3.3 years and in 17% of men over 2.7 years. Each 1 SD increase in baseline BMAT was associated with a 3.9 mg²/cm⁴/year greater loss of spine compressive strength index (p value = .003), a 0.9 mg/cm³/year greater loss of spine trabecular BMD (p value = .02), and a 1.2 mg/cm³/year greater loss of femoral neck trabecular BMD (p value = .02) in women. Among men, there were no associations between BMAT and changes in bone density/strength. There were no associations between BMAT and incident fractures in women or men. In conclusion, we found greater BMAT is associated with greater loss of trabecular bone at the spine and femoral neck, and greater loss of spine compressive strength, in older women.