Animal studies demonstrated that phytoestrogen had a protective effect against bone loss after ovariectomy. However, data on dietary phytoestrogen intake as well as its relationship with bone mineral density (BMD) in human are not available. Six hundred fifty southern Chinese women, aged 19 to 86 yr, were recruited to determine their dietary phytoestrogen intake by a food frequency questionnaire. BMDs at the lumbar spine and hip region were measured using dual energy x-ray absorptiometry. The subjects were analyzed according to various tertiles of phytoestrogen intake. Among the postmenopausal women (n = 357), significant differences in the lumbar spine (L2-4) BMD (0.820 +/- 0.145 vs. 0.771 +/- 0.131 g/cm2, P < 0.05) and Ward's triangle BMD (0.450 +/- 0.151 vs. 0.415 +/- 0.142 g/cm2; P < 0.05) were found between the highest and lowest intake of isoflavone after adjusting for age, height, weight, years since menopause, smoking, alcohol consumption, HRT usage, and daily calcium intake. Women with the highest intake of isoflavone had significantly lower levels of serum PTH (19.38 +/- 14.61 vs. 26.56 +/- 11.19 pg/ml; P < 0.05), osteocalcin (4.95 +/- 3.61 vs. 6.69 +/- 5.05 mg/liter; P = 0.05), and urinary N-telopeptide (34.18 +/- 25.31 vs. 49.66 +/- 41.00 nmol bone collagen equivalents/mmol creatinine; P < 0.05) when compared with those with the lowest intake of isoflavone. No association between dietary phytoestrogen intake and BMDs was seen in the premenopausal women with high endogenous E (n = 293). In conclusion, postmenopausal women with habitually high intake of dietary isoflavone are associated with higher BMD values at both the spine and hip region. Customarily high isoflavone intake may help to reverse the state of secondary hyperparathyroidism associated with E withdrawal and hence lower the rate of bone turnover in postmenopausal women.