Better methods to predict prognosis can play a supplementary role in administering individualized treatment for breast cancer patients. Altered expressions of PTHrP and TGF-β have been observed in various types of human cancers. The objective of the current study was to evaluate the association of PTHrP and TGF-β level with the clinicopathological features of the breast cancer patients.Immunohistochemistry was used to examine PTHrP and TGF-β protein expression in 497 cases of early breast cancer, and Kaplan-Meier method and COX's Proportional Hazard Model were applied to the prognostic value of PTHrP and TGF-β expression.Both over-expressed TGF-β and PTHrP were correlated with the tumor in larger size, higher proportion of axillary lymph node metastasis and later clinical stage. Additionally, the tumors with a high TGF-β level developed poor differentiation, and only TGF-β expression was associated with disease-free survival (DFS) of the breast cancer patients. Followed up for a median of 48 months, it was found that only the patients with negative TGF-β expression had longer DFS (P < 0.05, log-rank test). Nevertheless, those with higher PTHrP expression tended to show a higher rate of bone metastasis (67.6 % vs. 45.8 %, P = 0.019). In ER negative subgroup, those who developed PTHrP positive expression presented poor prognosis (P < 0.05, log-rank test). The patients with both positive TGF-β and PTHrP expression were significantly associated with the high risk of metastases. As indicated by Cox's regression analysis, TGF-β expression and the high proportion of axillary lymph node metastasis served as significant independent predictors for breast cancer recurrence.TGF-β and PTHrP were confirmed to be involved in regulating the malignant progression in breast cancer, and PTHrP expression, to be associated with bone metastasis as a potential prognostic marker in ER negative breast cancer.