Associations between caffeine and coffee consumption and breast cancer risk are uncertain, with studies suggesting inverse and null associations. Variation in cytochrome P450 1A2 (CYP1A2), a gene responsible for caffeine metabolism, may modify these associations. Cases (n = 3,062) were recruited through the Ontario Cancer Registry and controls (n = 3,427) through random digit dialing. Logistic regression was used to evaluate associations between breast cancer risk and intakes of 7 caffeine-containing items and total caffeine, and examine whether a genetic variant in CYP1A2 (rs762551) modified these associations. Analyses were stratified by estrogen receptor (ER), menopausal, and smoking status. Generally, coffee and caffeine were not associated with breast cancer risk; however, a significant reduction in risk was observed with the highest category of coffee consumption [≥5 cups per day vs. never, multivariate-adjusted odds ratio (MVOR) = 0.71, 95% confidence interval (CI): 0.51, 0.98]. Variant rs762551 did not modify associations. In stratified analyses, high coffee intake was associated with reduced risk of ER- (MVOR = 0.41, 95% CI: 0.19, 0.92) and postmenopausal breast cancer (MVOR = 0.63, 95% CI: 0.43, 0.94). High coffee consumption, but not total caffeine, may be associated with reduced risk of ER- and postmenopausal breast cancers, independent of CYP1A2 genotype. Further studies are needed to replicate these findings.