Objective To evaluate the association of known polymorphisms in the lipid metabolic pathway with body mass index (BMI), and estimate their interactions with soybean food intake. Methods A community-based cross-sectional survey was conducted in a Chinese Han population. BMI, soybean food intake, and single nucleotide polymorphisms of rs599839, rs3846662, rs3846663, rs12916, rs174547, rs174570, rs4938303, and rs1558861 were measured in 944 subjects. A multivariate logistic regression was used to analyze the association of the studied polymorphisms with BMIs. The expectation-maximization algorithm was employed to evaluate the extent of linkage disequilibrium between pairwise polymorphisms. The gene-environment interaction was assessed in the general multifactor dimensionality reduction model. Results The polymorphisms of rs3846662 and rs3846663 were associated with 10% highest BMIs when comparing to the 10% lowest values both in individuals and haplotype-based association tests. Although no statistically significant gene-environment interactions were found, people with the haplotype composed of C allele in rs3846662 and T allele in rs3846663 and low frequency of soybean intake had significantly higher risk to overweight and obesity as compared with those with the haplotype consisting of T allele in rs3846662 and C allele in rs3846663 and highly frequent soybean food intake, with an odds ratio of 1.64 (95%confidence interval: 1.15-2.34, P