Repressive histone tail modifications have been associated with molecular breast cancer subtypes. We investigated whether histone 3 lysine 9 trimethylation (H3K9me3) and histone 3 lysine 27 trimethylation (H3K27me3) were associated with tumor features and subtypes while adjusting for prospectively collected reproductive and lifestyle breast cancer risk factors. We have tissue microarray data with immunohistochemical marker information on 804 incident cases of invasive breast cancer diagnosed from 1976-2000 in the Nurses' Health Study. Tissue microarray sections were stained for global H3K9me3 and H3K27me3, and scored into four categories. Multivariate odds ratios (OR) and 95 % confidence intervals (CI) were calculated using logistic regression models for tumor features and subtypes, adjusting for breast cancer risk factors. While there were no significant associations between H3K9me3 and tumor features, H3K27me3 was significantly associated with lower grade tumors compared to high grade tumors in the multivariate model (OR = 1.95, 95 % CI 1.35-2.81, p = 0.0004). H3K27me3 was suggestively associated with estrogen receptor-positive (ER+) tumors (OR = 1.47, 95 % CI 0.97-2.23, p = 0.07). In subtype analyses, H3K27me3 was positively associated with the luminal A subtype compared to all other subtypes (OR = 1.42, 95 % CI 1.14-1.77, p = 0.002), and was inversely associated with HER2-type (OR = 0.58, 95 % CI 0.37-0.91, p = 0.02) and basal-like breast cancer (OR = 0.52, 95 % CI 0.36-0.76, p = 0.0006). In the largest immunohistochemical examination of H3K9me3 and H3K27me3 in breast cancer, we found that H3K27me3 positivity, but not H3K9me3, was associated with lower grade tumors and the luminal A subtype after adjusting for reproductive and lifestyle breast cancer risk factors.