Backgroundβ-thalassemia (β-TH) is a hereditary hemolytic anemia that results in deficient hemoglobin (Hb) synthesis. It is characterized by ineffective erythropoiesis, anemia, splenomegaly, and systemic iron overload. Exploration new potential biomarkers and drug candidates is important to facilitate the prevention and treatment of β-TH.MethodsWe applied quasi-targeted metabolomics between wild type (Wt) and heterozygous β-TH mice (Th3/+), a model of non-transfusion-dependent β-TH intermedia, in plasma and peripheral blood (PB) cells. Futher data was deeply mined by Kyoto Encyclopedia of Genomes (KEGG) and machine algorithms methods.ResultsUsing KEGG enrichment analysis, we found that taurine and hypotaurine metabolism disorders in plasma and alanine, aspartate and glutamate metabolism disorders in PB cells. After systematically anatomize the metabolites by machine algorithms, we confirmed that alpha-muricholic acidUP and N-acetyl-DL-phenylalanineUP in plasma and Dl-3-hydroxynorvalineUP, O-acetyl-L-serineUP, H-abu-OHUP, S-(Methyl) glutathioneUP, sepiapterinDOWN, and imidazoleacetic acidDOWN in PB cells play key roles in predicting the occurrence of β-TH. Furthermore, Sepiapterin, Imidazoleacetic acid, Methyl alpha-D-glucopyranoside and alpha-ketoglutaric acid have a good binding capacity to hemoglobin E through molecular docking and are considered to be potential drug candidates for β-TH.ConclusionThose results may help in identify useful molecular targets in the diagnosis and treatment of β-TH and lays a strong foundation for further research.