Objective To mine the ferroptosis-related genes associated with breast cancer prognosis through The Cancer Genome Atlas (TCGA) database, and construct a prognosis model.Methods The transcription group and clinical data were downloaded from TCGA database, differentially expressed ferroptosis-related genes in tumor tissues and adjacent normal tissues related to prognosis were obtained, and risk score model was constructed by least absolute shrinkage and selection operator (LASSO) regression. The patient information obtained from TCGA database was used as the model test set data. The effectiveness of the model was evaluated by receiver operating characteristic (ROC) curve, and univariate and multivariate Cox regression analyses were used to evaluate whether the differentially expressed ferroptosis-related genes and the risk scores could be used as prognostic factors. The model was verified by International Cancer Genome Consortium (ICGC) and Gene Expression Omnibus (GEO) databases.Results A total of 51 ferroptosis-related genes, which were differentially expressed in breast cancer tissues and adjacent normal tissues, were screened out. Univariate Cox regression analysis showed that 11 of them were associated with prognosis. A prognostic risk score model (risk score＝ALOX15×0.11＋CHAC1×0.07＋CISD1×0.15＋ CS×0.24＋GCLC×0.04＋GPX4×［ －0.07］＋NCOA4×0.17＋EMC2×0.30＋G6PD×0.19＋ACSF2×［ －0.04］＋ SQLE×0.12) for breast cancer was constructed with the 11 genes. ROC curve analysis showed that the area under curve (AUC) of the model in predicting the 2-, 4- and 6-year survival rates of breast cancer patients in the test set were 0.678, 0.680 and 0.612, respectively. The results of multivariate Cox regression analysis showed that the risk score could be used as an independent predictor (hazard ratio＝3.104, P＜0.001). The patients were divided into high-risk group (risk score≥4.277) and low-risk group (risk score＜4.277) according to the risk scores of the prognosis model. In the test set and validation set, the survival rate of highrisk patients was significantly lower than that of low-risk patients (both P＜0.001).Conclusion The prognosis model of breast cancer based on ferroptosis-related genes has better predictive performance. The ferroptosis-related genes in this model provides new targets for targeted therapy of breast cancer.