The Golgi apparatus (GA), an organelle that processes, sorts, and transports proteins synthesized by the endoplasmic reticulum, is also involved in many cellular processes associated with cancer, such as angiogenesis, the innate immune response, and tumor invasion and migration. We aimed to construct a breast cancer (BC) prognosis prediction model based on GA-related genetic information to evaluate the prognosis of patients with BC more accurately than existing models and to stratify patients for clinical therapy. In this study, The Cancer Genome Atlas-breast invasive carcinoma was used as the training cohort, and the Molecular Taxonomy of Breast Cancer International Consortium cohort was used as the validation cohort. Using bioinformatics methods, we constructed a GA-related gene risk score (GRS). The GRS was used to divide BC patients into a high-GRS group and a low-GRS group, and functional analysis, survival analysis, mutation analysis, immune landscape analysis, and metabolic analysis were performed to compare the 2 groups. Finally, a nomogram was constructed for clinical application. The genes in the GRS model were mainly related to the glucose metabolism pathway, and the main mutations in the 2 groups of patients were mutations in TP53 and CHD1. The mutation rate in the high-GRS group was greater than that in the low-GRS group. The high GRS group had higher tumor immune activity glycolysis; the pentose phosphate pathway tended to be the dominant metabolic pathways in this group, while fatty acid oxidation and glutamine catabolism tended to be dominant in the low-GRS group. GA-related genes were used to construct a prediction model for BC patients and had high accuracy in predicting prognosis. The mutations associated with the GRS are mainly TP53 and CDH1. Interestingly, the GRS is correlated with glucose metabolism in terms of gene expression and functional enrichment. In summary, the role of GRS-related genes in glucose metabolism is worthy of further study.