Endogenous estrogen and tobacco smoke are putative risk factors for breast cancer. The genes CYP1A1 and 1B1 are involved in the metabolism of endogenous estrogen as well as of tobacco specific compounds. In a German population based case-control study of premenopausal breast cancer risk with 394 cases and 838 controls, the effects of genetic polymorphisms of theses metabolising genes on the association between estimated estrogen exposures, as well as active/passive tobacco exposure and breast cancer risk were investigated. The alleles CYP1A1*1, *2A, *2B, *4 and CYP1B1*1, *2, *3, *4, alone or in combination, were not apparently associated with breast cancer risk. However, CYP1B1*3 modified the association between estimated estrogen exposure and cancer risk, i.e. for higher estrogen compared to lower estrogen exposure, odds ratios increased with increasing number of *3 alleles from 0.6(0.3-1.2) to 0.9(0.6-1.4) to 2.0(1.0-3.9), (p for multiplicative interaction: 0.015). Overall, active and passive smoking increased breast cancer risk. Different polymorphisms slightly modified this association, although all tests for multiplicative interaction between genotype and smoking variables were non-significant. Active and passive smoking were associated with highly increased risks for carriers of at least one CYP1B1*2 allele, but with risks around unity for carriers of at least one CYP1B1*4 allele. In carriers of at least one variant CYP1A1 allele, tobacco exposure was not associated with breast cancer risk, whereas in those homozygous for the wildtype allele smoking increased risk. These findings should be interpreted with caution because of small sample size. The results suggest that 1) CYP1B1 genotype has a differential effect on the association between estrogen exposure and breast cancer and 2) CYP1B1 as well as CYP1A1 genotype seem to modify only moderately, if at all, the association between smoking variables and breast cancer risk.