Background Overexpression of the HER2 proto-oncogene in human cancer cells has been associated with a poor prognosis, and survival improves with therapy targeting the HER2 gene. Animal studies and protein modeling suggest that the Ile655Val polymorphism located in the transmembrane domain of the HER2 protein might influence breast cancer development by altering the efficiency of homodimerization. Methods To investigate this genetic polymorphism, incident cases of invasive breast cancer (N = 1,094) and population controls of a similar age (N = 976) were interviewed during 2001 to 2003 regarding their risk factors for breast cancer. By using DNA collected from buccal samples mailed by the participants, the HER2 Ile655Val polymorphism was evaluated with the Applied Biosystems allelic discrimination assay. Odds ratios (ORs) and 95% confidence intervals (95% CIs) were estimated by logistic regression adjusted for numerous breast cancer risk factors. Analysis was restricted to women with self-reported European descent. Results Prevalence of the Val/Val genotype was 5.6% in cases and 7.1% in controls. In comparison with the Ile/Ile genotype, the Ile/Val genotype was not significantly associated with breast cancer risk (OR 0.97, 95% CI 0.79 to 1.18), whereas the Val/Val genotype was associated with a reduced risk (OR 0.63, 95% CI 0.42 to 0.92). This inverse association seemed strongest in older women (OR 0.51, 95% CI 0.29 to 0.89 for women aged more than 55 years), women without a family history of breast cancer (OR 0.54, 95% CI 0.35 to 0.84), postmenopausal women with greater body mass index (OR 0.43, 95% CI 0.20 to 0.91 for a body mass index of 25.3 kg/m2 or more), and cases diagnosed with non-localized breast cancer (OR 0.49, 95% CI 0.26 to 0.90). Conclusion Although results from our population-based case-control study show an inverse association between the HER2 Ile655Val polymorphism and risk of invasive breast cancer, most other studies of this single-nucleotide polymorphism suggest an overall null association. Any further study of this polymorphism should involve sample populations with complete risk factor information and sufficient power to evaluate gene-environment interactions between the HER2 polymorphism and factors such as age and family history of breast cancer.