Recent interest has been shown in the potential beneficial effects of phytoestrogens on bone health. As the early years of menopause are a period of rapid bone loss, and the risk for osteoporosis increases substantially, the habitual intake of soy protein and isoflavones may play a role in the retardation of bone loss. This paper reports the results of the baseline cross-sectional analysis of the association between dietary soy protein intake and bone mineral density/content in a population-based study of Chinese women. The sample comprised 454 healthy Chinese women (mean age 55.1 +/- 3.57) within the first 12 years of postmenopause. We estimated the dietary intake of soy protein and isoflavones, and other key nutrients, including dietary protein and calcium, using the quantitative food frequency method. Bone mineral density (BMD) and content (BMC) at the spine, hip and total body were measured with a dual energy X-ray densitometer (Hologic 4500A). Soy protein consumption was categorized as quartiles of intake, and related to BMD values at the spine and hip, and BMC of total body. Stratified analyses were carried out among women within or at least 4 years postmenopausal. We observed few differences in BMD/BMC values among the intake quartiles in women within the first 4 years of menopause. However, among the later postmenopausal women, we noted a dose-response relationship with increasing higher BMD values at the trochanter, intertrochanter as well as the total hip and total body with increasing soy protein intake quartiles (P<0.05 from tests for trend). The BMD values differed by about 4-8% between the first and fourth soy protein intake quartiles. Though women from the fourth intake quartile had a 2.9% higher BMD value compared with those from the first intake quartile, the difference was not statistically significant. Stepwise multiple linear regression analyses showed the association between soy intake quartiles and hip BMD as well as total body BMC values remained after adjusting for body weight, which was retained in the final model. Analyses based on soy isoflavones content yielded similar results. This study demonstrated that, among women after the initial few years postmenopausal, soy protein/isoflavones intake had a modest but significant association with hip BMD as well as total body BMC. The effects of soy protein and soy isoflavones on bone health should be further explored in populations with habitual dietary soy intake.