An ultra-performance liquid chromatography-quadrupole-time of flight mass spectrometric（UPLC/Q-TOF MS） method was developed for the study on metabolic variations of urine from diabetes mellitus patients and healthy volunteers.PLS-DA was used for group differentiation and potential biomarkers selection.As shown in the scores plot,the distinct clustering among diabetes mellitus patients and health controls was observed.In carbohydrate metabolites,the sodium glucuronic acid content of the patients was significantly higher than that of the normal control group.Compared with the normal control group,the fatty acids contents of the patients were higher than those of the control group,contents of 20-carboxy-arachidonic acid,palmitic acid in the urine of patients with diabetes were higher than that in normal subjects.The metabolite contents of amino acids including tyrosine,N-beta-alanyl-L histidine and tryptophan of the patients were lower than those of the normal,but the contents of glutaric acid,nicotinoyloxy glycine,aspartic acid,phenylalanine,etc.of the patients were lower than those of the normal people.These metabolites are associated with glucose and lipid regulating substances.This study showed the potential of metabolomic strategies for explanation of the mechanism of diseases and detection of specific biomarkers.