<p>The impact of oxidative stress on breast cancer remains unknown, and further research is urgently needed. Datasets from TCGA, METABRIC, and GEO were used in this study, and the oxidative stress scores were determined using the least absolute shrinkage and selection operator (LASSO) and Cox regression method. A prognostic model was established using the TCGA-BRCA dataset and verified using the METABRIC and GEO datasets. The microenvironment of breast cancer was evaluated using single-cell data; cell types and oxidative stress scores were evaluated in normal cells, pre-neoplastic cells, and breast cancer cells. An oxidative stress score was calculated with eight genes using LASSO and Cox regression analyses, and a prognostic model with three parameters—age, tumor stage, and oxidative risk score—was established. Good performance was demonstrated by prognostic analysis. Immune cell infiltration indicated that endothelial cells and cancer-associated fibroblasts were higher in the high oxidative stress score group. Single-cell analysis showed that endothelial cells have the lowest degree of differentiation in breast cancer, and was confirmed by pseudo-time analysis. Gene ontology analysis proved that oxidative stress participates in the tumorigenesis and metastasis of breast cancer, which indicate that oxidative stress impact on endothelial cell and participate in tumorigenesis of breast cancer. Interestingly, the oxidative stress score was lower in breast cancer cells because their endothelial cells overexpressed antioxidant-related genes, which allowed breast cancer cells to escape oxidative stress-induced apoptosis. Although oxidative stress is involved in tumorigenesis, it is prevented and eliminated during the promotion and progression stages. Better understanding of tumor microenvironment may facilitate the identification of therapeutic targets and precise treatment.</p>