Neonatal cholestatic disorders are a group of hepatobiliary diseases occurring in the first 3 months of life. The most common causes of neonatal cholestasis are infantile hepatitis syndrome (IHS) and biliary atresia (BA). The clinical manifestations of the two diseases are too similar to distinguish them. However, early detection is very important in improving the clinical outcome of BA. Currently, a liver biopsy is the only proven and effective method used to differentially diagnose these two similar diseases in the clinic. However, this method is invasive. Therefore, sensitive and non-invasive biomarkers are needed to effectively differentiate between BA and IHS. We hypothesized that urinary metabolomics can produce unique metabolite profiles for BA and IHS. The aim of this study was to characterize urinary metabolomic profiles in infants with BA and IHS, and to identify differences among infants with BA, IHS, and normal controls (NC). Urine samples along with patient characteristics were obtained from 25 BA, 38 IHS, and 38 NC infants. A non-targeted gas chromatography-mass spectrometry (GC-MS) metabolomics method was used in conjunction with orthogonal partial least squares discriminant analysis (OPLS-DA) to explore the metabolomic profiles of BA, IHS, and NC infants. In total, 41 differentially expressed metabolites between BA vs. NC, IHS vs. NC, and BA vs. IHS were identified. N-acetyl-d-mannosamine and alpha-aminoadipic acid were found to be highly accurate at distinguishing between BA and IHS. BA and IHS infants have specific urinary metabolomic profiles. The results of our study underscore the clinical potential of metabolomic profiling to uncover metabolic changes that could be used to discriminate BA from IHS.