Type 2 diabetes mellitus (T2DM) is a complex metabolic disorder frequently accompanied by cognitive dysfunction and affects nearly 30% of people with diabetes, is an independent risk factor for cognitive dysfunction. Exposure to high-altitudes (above 2500 meters above sea level) with hypobaric hypoxia can also lead to cognitive dysfunction. Which is also a risk factor for cognitive dysfunction. Therefore, the two risk factors of diabetes and high-altitude combined, the damage to cognitive dysfunction may be more serious, and may even develop into dementia. So, early diagnosis and discovery of cognitive function biomarkers of diabetes at high-altitude are of great significance for prevention and treatment. This study is to investigate the early specific metabolites biomarkers of diabetic cognitive dysfunction in high-altitude by using metabolomics technology. Total 400 subjects were divided into four groups, diabetics in high-altitude (H-T2DM), normal control in high-altitude (H-HC), diabetics in low-altitude (L-T2DM),normal control in low-altitude (L-HC). Cognitive deficits were assessed in H-T2DM and L-T2DM using a cognitive function recognition test. The recognition test showed significant cognitive impairment in the H-T2DM. Serological results showed higher hemoglobin (HbA1c) values in the H-T2DM. Four groups of serum samples were analyzed by ultra performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UPLC-Q/TOF-MS) platform. The stability of the model was verified by principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA). Metabolites with significant differences were screened out as potential biomarkers through the OPLS-DA model according to the importance of variables in the project (VIP >1) and the P value of the t-test (P<0.05). Through multivariate statistical and integrated analysis, a total of 26 differentially expressed endogenous metabolites were identified (18 up-regulated and 8 down-regulated in H-T2DM). Through pathway topology analysis, we found that the pipecolic acid, lauric acid, guanosine and kaempferol could be accepted as early biomarkers of diabetic cognitive impairment in high-altitude. The prediction accuracy rate was as high as 92%. The identified biomarkers are mainly related to lysine degradation, fatty acid biosynthesis, purine metabolism and metabolic pathways. Through the verification of multi-center population, it was found that guanosine is the biomarker with the most potential to become an early biomarker. This study reveals for the first time reliable biomarkers for early diagnosis of diabetes at high-altitude. It may be provide new ideas and strategies for early diagnosis.