(1)H nuclear magnetic resonance (NMR) based metabonomic method was used to characterize the profile of low-molecular-weight endogenous metabolites in mouse (Mus musculus) serum following exposure to Taihu Lake source water for 90 days. The (1)H NMR spectra of mice sera were recoded and a total of 21 metabolites were identified. Data reduction and latent biomarkers identification were processed by pattern recognition (PR) analysis. The principal component analysis (PCA) and partial least square discriminant analysis (PLS-DA) identified differences in metabolic profiles between control and treatment groups. A number of serum metabolic perturbations were observed in sera of source water treatment mice compared to control mice, including decreased levels of acetone, pyruvate, glutamine, lysine and citrate. These results indicated that Taihu Lake source water could induce energy metabolism changes in mouse related to fatty acid β-oxidation, tricarboxylic acid (TCA) cycle, citric acid cycle, and metabolism of some amino acids. (1)H NMR-based metabolic profiling provides new insight into the toxic effect of Taihu Lake source water, and suggests potential biomarkers for noninvasive monitoring of health risk.