Predictive metabolic biomarkers for the recurrent luminal breast cancer (BC) with hormone receptor (HR)-positive and human epidermal growth factor receptor type 2 (HER2)-negative are lacking. High levels of O-GlcNAcylation (O-GlcNAc) and pyruvate kinase isoenzyme M2 (PKM2) are associated with malignancy in BC; however, the association with the recurrence risk remains unclear. We first conduct survival analysis by using the METABRIC dataset to assess the correlation of PKM2 expression with BC clinical outcomes. Next, patients with HR+/HER2-luminal BC were recruited for PKM2/O-GlcNAc testing. Logistic regression and receiver operating characteristic curve analysis were performed to evaluate the 10-year DFS predicted outcome. Survival analysis of the METABRIC dataset revealed that high expression of PKM2 was significantly associated with worse overall survival in luminal BC. The high expression of O-GlcNAc or PKM2 was a significant independent marker for poor 10-year DFS using immunohistochemical analysis. The PKM2 or O-GlcNAc status was a significant predictor of DFS, with the combination of PKM2–O-GlcNAc status and T stage greatly enhancing the predictive outcome potential. In summary, O-GlcNAc, PKM2, and T stage serve as good prognostic discriminators in HR+/HER2− luminal BC.