Objective To analyse the prognostic factors and 5 prognostic indices in breast cancer patients with brain metastases after whole-brain radiotherapy (WBRT). Methods Log-rank test (for univariate prognostic analysis) and Cox regression analysis (for multivariate prognostic analysis) were performed in 99 breast cancer patients with BM who received WBRT from January 2006 to February 2010. The sensitivities and specificities of prognostic indices, including recursive partitioning analysis (RPA), graded prognostic assessment (GPA), basic score for brain metastases (BSBM), Rades, and Carsten, were calculated and compared. Results The median follow-up was 49 months;the median survival was 10 months. The univariate analysis showed that age, Karnofsky performance score (KPS) at the diagnosis of BM, presence of extracranial metastases, primary tumor control, number of BM, and systemic treatment after WBRT were significantly correlated with overall survival (χ²=0.00—55.51, P=0.000—0.013). The multivariate analysis showed that KPS of<70 and systemic treatment after WBRT were significantly correlated with overall survival (χ²=35.26, P=0.000;χ²=7.21, P=0.007). In predicting ≤3 months of survival for breast cancer patients with BM, the sensitivities and specificities of RPA, GPA, BSBM, Rades, and Carsten were 100% and 85%, 95% and 62%, 95% and 86%, 95% and 84%, and 95% and 85%, respectively. Conclusions Systemic treatment after WBRT can improve overall survival in breast cancer patients with BM. In predicting ≤3 months of survival, RPA is the most sensitive index, and BSBM is the most specific index.