The primary objective of this study was to determine and characterize surrogate biomarkers that can predict nephrotoxicity induced by mercuric chloride (HgCl₂) using urinary proton nuclear magnetic resonance (¹H NMR) spectral data. A procedure for (1)H NMR urinalysis using pattern recognition was proposed to evaluate nephrotoxicity induced by HgCl₂ in Sprague-Dawley rats. HgCl₂ at 0.1 or 0.75 mg/kg was administered intraperitoneally (i.p.), and urine was collected every 24 h for 6 days. Animals (n=6 per group) were sacrificed 3 or 6 days post-dosing in order to perform clinical blood chemistry tests and histopathologic examinations. Urinary ¹H NMR spectroscopy revealed apparent differential clustering between the control and HgCl₂ treatment groups as evidenced by principal component analysis (PCA) and partial least square (PLS)-discriminant analysis (DA). Time- and dose-dependent separation of HgCl₂-treated animals from controls was observed by PCA of ¹H NMR spectral data. In HgCl₂-treated rats, the concentrations of endogenous urinary metabolites of glucose, acetate, alanine, lactate, succinate, and ethanol were significantly increased, whereas the concentrations of 2-oxoglutarate, allantoin, citrate, formate, taurine, and hippurate were significantly decreased. These endogenous metabolites were selected as putative biomarkers for HgCl₂-induced nephrotoxicity. A dose response was observed in concentrations of lactate, acetate, succinate, and ethanol, where severe disruption of the concentrations of 2-oxoglutarate, citrate, formate, glucose, and taurine was observed at the higher dose (0.75 mg/kg) of HgCl₂. Correlation of urinary (1)H NMR PLS-DA data with renal histopathologic changes suggests that ¹H NMR urinalysis can be used to predict or screen for HgCl₂-induced nephrotoxicity.