Osteoporosis is one of the consequences of aging, and it remains underdiagnosed and undertreated; this study aimed to present the characteristics and prevalence of osteoporosis in elderly men by conducting a nationwide survey in Taiwan. The participants were enrolled between 2008 and 2011, and bone mineral density (BMD) was measured via dual-energy X-ray absorptiometry for the hip (total), lumbar spine (L1-4), and femoral neck (FN). Patients with rheumatoid arthritis, female patients, and those using steroids were excluded. Osteoporosis was defined as a T-score at the FN of ≤-2.5. This study included 3734 men of mean age 70.0 ± 9.3 years, accounting for the prevalence of osteoporosis at 9.7%. Participants with osteoporosis had a significantly older age, lower body weight, shorter height and more previous fractures than those without osteoporosis. The mean BMD at FN was 0.534 ± 0.056 and 0.791 ± 0.115 (g/cm2 ) in participants with and without osteoporosis, respectively (P < 0.001). The FN and hip (total) BMD showed a significant negative correlation with age (r = -0.234, P < 0.001) and (r = -0.003, P < 0.001), respectively, but not at L1-4 (r = 0.00, P = 0.540). A history of fracture is the most important risk factor associated with male osteoporosis (odds ratio, 2.50; 95% CI, 1.49-4.21; P = 0.006). The associated factors for male osteoporosis are aging, lower body weight, and a history of fracture; the BMDs at FN and hip (total), but not L1-4, are inversely correlated with age. We recommend that BMD at the proximal femur be the preferred site to evaluate osteoporosis for elderly male subjects. © 2018 Asia Pacific League of Associations for Rheumatology and John Wiley & Sons Australia, Ltd.