We investigated the prognostic influence of androgen receptor (AR) on breast cancer. AR status was assessed using immunohistochemistry with tissue microarrays from 395 operable primary breast cancer patients who received curative surgery. The Kaplan–Meier estimator was used to analyze the survival rates and a log-rank test was used to determine the significance of the differences in survival. The Cox proportional hazards model was used to calculate the hazard ratio (HR) and the 95% confidence interval (CI) of survival. There were 203 (51.4%) subjects with a low expression of AR, and 192 patients (48.6%) with a high expression rate. The high AR expression group showed superior overall survival (p = 0.047) and disease-free survival (p = 0.004) when compared with the low AR expression group. The high AR expression group showed superior systemic recurrence-free survival when compared with the low AR expression group (p = 0.027). AR was an independent prognostic factor for both overall survival (HR, 0.586; 95% CI, 0.381–0.901; p = 0.015) and disease-free survival (HR, 0.430; 95% CI, 0.274–0.674; p < 0.001). A high AR expression was a significant favorable prognostic factor only in the subgroups with positive hormone receptors (HRc) and negative human epidermal growth factor receptor 2 (HER2) when considering disease-free survival (p = 0.026). The high AR expression group was significantly associated with superior overall survival and disease-free survival when compared with the low AR expression group with breast cancer patients. AR was a significant independent prognostic factor for both overall survival and disease-free survival. The prognostic impact of AR was valid in the HRc(+)/HER2(−) subtype when considering disease-free survival. These findings suggest the clinical usefulness of AR as a prognostic marker of breast cancer in clinical settings.