Abstract: OBJECTIVE In the present study， we aimed to investigate the metabolite changes in the brain tissues and serum from middle cerebral artery occlusion (MCAO) model rats using gas chromatography-mass spectrometry (GC-MS)-based metabolomics technique. METHODS The transient model of middle cerebral artery occlusion (MCAO) rats was established using Zea Longa thread embolism methods. GC-MS-based metabolomics technique were performed to investigate the metabolites in brain tissues and serum from MCAO and Sham group rats. OmicShare tools were to perform the heatmap and hierarchical cluster analysis of the metabolites. Principal component analysis (PCA) were used to detect the difference between both group rats. A combination of VIP value of Partial least squares discriminant analysis (PLS-DA) with p value of t test were to screen differentially expressed metabolites. The pathway database MetaboAnalyst were applied to analyse the pathway that these differentially expressed metabolites participated in. RESULTS We identified and quantified 95 metabolites in brain tissues and 105 in serum， respectively. Compared with the sham group， the concentrations of urea， valine， leucine， isoleucine， mannose， urea， phenylalanine， and methionine in serum from MCAO group rats were significantly increased（P<0.05）. However， the concentrations of proline， alanine， glutamic acid， myo-inositol， serine， glycerol， linoleic acid， and 4-hydroxyproline were markedly decreased（P<0.05）. In the brain tissues from MCAO group rats， the concentrations of alanine， lactic acid， urea， leucine， proline， valine， 2-hydroxyisobutyric acid， and isoleucine were significantly increased（P<0.05）. The metabolic pathway results showed that these metabolites were associated wih some amino acid biosynthesis and metabolism， linoleic acid metabolism， methane metabolism， glycerolipid metabolism， and aminoacyl-tRNA biosynthesis， etc. CONCLUSION Our findings demonstrate the presence of characteristic changes in metabolites in the brain tissues and serum from MCAO rats and highlights the utility of metabolomics in identifying elements of stroke disease pathogenesis and for the development of early diagnostic biomarker of the disease.