Long-term precision, as well as reproducibility, is important for monitoring bone mineral density (BMD) alteration in response to aging or therapy. In order to investigate which bone densitometry and which skeletal site are clinically useful for monitoring bone mass, we examined the standardized long-term precision of several bone density measurements in 83 healthy Japanese women. Annual BMD measurements were performed for 5 or 6 years using dual X-ray absorptiometry (DXA) on the lumbar spine, radius (EXP5000) and calcaneus (HeelScan); peripheral quantitative computed tomography (pQCT) on the radius (Densiscan1000); and quantitative ultrasound (QUS) on the calcaneus (Achilles+). The long-term precision error for the individual subject was given by the standard error of estimate (SEE), and the standardized long-term precision was defined as the percentage coefficient of variation (CV%) divided by the percentage ratio of the annual bone-loss rate. Based on the CV% of spinal DXA, speed of sound (SOS) and diaphyseal pQCT showed significantly higher precision than others, while radial ultradistal (UD) DXA and heel DXA showed significantly lower precision. The long-term precision errors of other measurements were statistically the same as that of the spinal DXA. The spinal DXA, the radial DXA, and pQCT at both the distal metaphysis and diaphysis showed high rates of annual bone loss. The radial trabecular BMD (pQCT) was significantly higher than that of spinal DXA. The annual rates of bone loss of QUS and of heel DXA were significantly lower than that of spinal DXA. Taken together, standardized long-term precision was obtained in the spinal DXA and radial pQCT. In conclusion, spinal DXA and radial pQCT were considered the most useful monitoring method for osteoporosis, while QUS was considered less useful.