The aim of this study was to describe a metabolomic study of breast cancer using 1 H-NMR combined with bioinformatics analysis. 1 H-NMR spectroscopy combined with multi-variate pattern recognition analysis was used to cluster the groups (serum and urine samples from breast cancer patients and healthy controls) and establish a breast-cancer-specific metabolites phenotype. Orthogonalpartial least-squares discriminant analysis (OPLS-DA) was capable of distinguishing serum and urine samples from breast cancer patients and healthy controls and establishing a breastcancer-specific metabolite profile. A total of 9 metabolites in serum concentration and 3 metabolites in urine concentration differed significantly between breast cancer patients and healthy controls. Serum samples from breast cancer patients were characterized by decreased concentrations of choline, glucose, histidine, valine, lysine, acetate, tyrosine and glutamic, accompanied by increased concentrations of lipid relative to healthy controls. In urine samples, the level of phenylacetylglycine and guanidoacetate was significantly lower, while the level of citrate was significantly higher in breast cancer patients relative to healthy controls. In conclusion, this study reveals the metabolic profile of serum and urine from breast cancer patients. NMRbased metabolomics has the potential to be developed into a novel clinical tool for diagnosis or therapeutic monitoring for breast cancer. However, because of limitations of methods and technique, further research and verification is needed.