Estrogen receptor (ER)-positive breast cancer is a distinct subpopulation of breast cancer exhibiting a high response rate to endocrine therapy. However, not all ER-positive patients respond to the therapy, and a subgrouping of ER-positive patients based on the physiology of estrogen signaling is expected to be useful for predicting the prognosis. This study has revealed that selected estrogen-regulated genes (ERGs) are useful in identification of a poor-prognosis population among ER-positive breast cancer patients. First, the expression levels of 11 ERGs, selected based on our earlier microarray study in cultured cells, were analyzed by means of real-time reverse transcription-PCR in 14 ER-positive human breast cancer tissues. The patients were clearly divided into two groups in cluster analysis. Then, we examined the expression levels of two representative ERGs, histone deacetylase 6 (HDAC6) and insulin-like growth factor binding protein 4 (IGFBP-4), in 62 ER-positive patients with immunohistochemistry to assess the impact of ERG expression on prognosis (median follow-up 4409 days). Positive HDAC6 staining was significantly correlated with a lower disease-free survival rate. Moreover, when the expression level of HDAC6 was assessed in combination with IGFBP-4 expression in the nucleus, the poor-prognosis patients were more accurately identified. This study has identified new candidate ERGs for prediction of prognosis, and we suggest that combined assessment of the expression levels of these ERGs will contribute to the clinically useful stratification of ER-positive breast cancer patients.