The effect of exercise practice on the radius bone mineral density (BMD) was investigated on 480 women at the age of perimenopause and later. The BMD at the 1/3 distal site of the radius on a non-dominant hand was measured using the Dual Energy X-ray Absorptiometry method. The subjects were divided into group E with regular exercising and group C without regular exercising. The difference of mean radius BMD values was not statistically significant between groups E and C. Each of the groups E and C was further classified into a perimenopause group and groups with stratified postmenopausal years (postmenopausal years less than 5, less than 10 and more than 10). Then, the mean radius BMD value of the perimenopause subgroup was taken as 100% for each of the groups E and C, and the relative radius BMD values of subgroups were compared between groups E and C. The relative BMD values for the subgroup of postmenopausal years more than 10 were higher in E group than those in C group, while no difference was observed in below postmenopausal 10 years between the two groups. To clarify the BMD difference among exercise events, group E was further divided into the following three subgroups: the subgroup practising events which give high-impact to arm (HI group), such as tennis, ping-pong, golf and volleyball, the subgroup practising swimming (SW group), and the subgroup practising events which give low-impact to arm (LI group), such as walking, running, aerobic dance and light gymnastics. No significant differences were observed in BMD values among the above sports event-stratified subgroups when postmenopausal years were less than 10 years, but the relative BMD values for the subgroup of postmenopausal years more than 10 were higher in HI and LI group than those in control group. These results suggest the possibility that exercise practice suppresses a postmenopausal decline in bone density for preventing osteoporosis from developing.