Dual-energy X-ray absorptiometry (DXA) was used to examine the effects of quantitative application of treadmill running exercise on bone mineral density (BMD) of the tibia and the fourth and fifth lumbar (L4 + L5) vertebrae in mature osteopenic rats. Twenty 37-week-old rats with bone loss, resulting from feeding a relatively low calcium diet for 14 weeks after ovariectomy at the age of 23 weeks, were divided into four groups of five rats each according to the intensity and duration of the exercise: 12 m/minute, 1 h/day in group EX1; 18 m/minute, 1 h/day in group EX2; 12 m/minute, 2 h/day in group EX3; and sedentary control in group CON. With a standard calcium diet, the exercise was performed 5 days a week for 12 weeks, and the BMD of both the right tibia and the L4 + L5 vertebrae was measured using DXA at weeks 0, 4, 8, and 12. At the end of 12 weeks of exercise, the right femur and the L5 vertebra were dissected and the mechanical strength was measured using a three-point bending test and a compression test, respectively. After 12 weeks of exercise, a significant increase in the tibial BMD was observed in only group EX1 compared with that in group CON (p = 0.0039, by two-way analysis of variance). However, any significant increase in the L4 + L5 vertebral BMD was not observed in any exercise groups compared with that in the control group. While a maximum breaking force of the femoral shaft in group EX1 was significantly greater than that in group CON (p < 0.05, by Mann-Whitney's U-test), that in groups EX2 and EX3 did not significantly differ from that in group CON. However, there was no significant difference in a maximum breaking force of the L5 vertebral body among all the exercise and control groups. These results indicated that the beneficial effects of treadmill running exercise under a standard calcium diet were recognized only in the weight-bearing bones of the mature osteopenic rats resulting from estrogen deficiency and inadequate calcium intake only when an optimal level of exercise was applied.