Tamoxifen is the most commonly used drug to treat estrogen receptor positive (ER+) breast cancer. However, many patients with ER+ breast cancer have experienced resistance and other adverse side effects following treatment with tamoxifen. Furthermore, clinical and pathological parameters have thus far failed to predict the efficiency of tamoxifen administration. Therefore, gene signature based models for the prediction of survival time of such patients are urgently needed. In the current study, gene expression levels and follow‑up information of samples from GSE17705 and GSE22219 databases were used to construct a risk score model based on Cox multivariate regression. The expression levels of 10 genes were included in the model: CCNB2, CCNA2, FOXD1, WSB2, RBPMS, CTDSP1, BIN3, SLBP, EPRS, FTO. The samples in the high‑risk group had a relative early distant relapse time period (median survival time of 3.75years) compared with the patients in the low risk group (median survival time of 6.5years, P<0.01). For further validation, a further two independent datasets (GSE26971, GSE58644) were assessed. The overall survival time period of patients with high‑risk scores in these datasets was significantly longer than those with low‑risk scores (P<0.01). Furthermore, the associations between clinical parameters and risk score were investigated, and it was revealed that the risk score was significantly correlated with tumor age, tumor stage and grade. In addition, a 5‑year survival nomogram was plotted in order to facilitate the utilization of risk score along with other clinical data. In summary, using the transcriptomic profile, a multi‑gene expression based risk score was developed and was revealed as being able to successfully predict the outcome of patients with ER+ breast cancer treated with tamoxifen for 5years.