Background: It has been reported that estrogen receptor beta ($ER{eta}$) mRNA expression was down-regulated during carcinogenesis and was inversely related to estrogen receptor alpha ($ER{alpha}$) expression in breast cancer. The association of $ER{eta}$ mRNA expression to tamoxifen resistance has also been reported. In this study, the expression of $ER{alpha}$ and $ER{eta}$ via immunohistochemistry (IHC) and reverse transcription-polymerase chain reaction (RT-PCR) was prompted, and an attempt was made to find out the relationship between $ER{eta}$ expression and recurrence in the hormonal therapy group, and between $ER{eta}$ expression and known prognostic factors. Methods: Tumor specimens were obtained at surgery from 67 female breast cancer patients during the period of September 1995 to December 2000. All the specimens were frozen in liquid nitrogen and kept at $-70^{circ}C$ until they were used. The medical records were analyzed retrospectively. The expressions of ER were analyzed using IHC and RT-PCR methods. Results: The median follow-up was at 93.0 months (range: 14-157 months). The percentage of $ER{alpha}+/ER{eta}+$, $ER{alpha}+/ER{eta}-$, $ER{alpha}-/ER{eta}+$, and $ER{alpha}-/ER{eta}$ group were 35.9% 9.4%, 47.2%, and 7.5%, respectively, in 53 patients with hormonal therapy. $ER{eta}$ was positive in 42 (82.3%) of 51 ER-positive patients. In the hormonal therapy group, the recurrence rates of each group was 15.8%, 0%, 40.0%, and 0%, respectively. In this group, the $ER{eta}$ expression tended to recur, but there was no clinical significance (p=0.084). Conclusion: The $ER{eta}$ expression may be a predictive marker of a poor response to endocrine therapy in breast cancer patients, although this needs to be confirmed in additional studies.