Dermatomyositis (DM) is a rare autoimmune myopathy characterized by skin lesions, proximal muscle weakness and muscle inflammation. The pathogenesis of DM is unclear, and identification of reliable biomarkers for early diagnosis of DM is critical for design of a specific therapy for this disease.To find and identify potential serum biomarkers in DM patients.We performed an untargeted metabolomic approach using UHPLC-MS/MS. The blood serum metabolomic profiles of 26 DM patients and 26 healthy controls were collected. Multivariate analysis of the metabolomic profile was applied to differentiate DM patients and controls and to find potential biomarkers.A significantly disturbed metabolic profile of DM patients was observed. Pathway analysis showed that aminoacyl-tRNA biosynthesis, phenylalanine, tyrosine and tryptophan biosynthesis, and nitrogen metabolism are the most prominently altered pathways in DM. Receiver operating characteristic curve indicated that glutamine, methionine, isoleucine, tryptophan, glutamic acid, indole, protocatechuic acid, and phenylalanine were potential biomarkers for DM diagnosis in terms of both sensitivity and specificity.Our study provides new insight into underlying mechanisms of DM, and we suggest that we should pay more attention to these metabolic pathways in the prevention and treatment of DM.