Objective In this work, the metabonomic methods utilizing LC-ESI-MS and multivariate statistical technique have been applied to investigate the metabolic profiles of hypoxic preconditioning of mice by the model of acute repetitive hypoxia.Methods The brain metabonomics of hypoxic preconditioning was compared with that of the control groups.Whole brain tissues were collected from both the normal mice and those mice suffered acute repetitive hypoxia.The brain sample data obtained with the LC-ESI-MS were analyzed by pattern recognition using principal component analysis.The human metabolome database and MS/MS data produced by the software of Mass Frontier were employed to identify and elucidate the metabolites of interests and some biomarkers.And their functions and biological significance were also analyzed.Results There were significantly different in the brain metabolic pattems between the hypoxia group and control group via the treatment of acute repetitive hypoxia.The major differences of the endogenous metabolite profiles in brain samples were observed between hypoxia group and control group.Phenylalanine, valine, glutamine,creatine, proline and spermidine were increased in brain hypoxia group; and phenylalanine, valine, glutamine, methionine and uracil were increased in heart hypoxia group, while glutamic acid was decreased both brain and heart.The biological significances of biomarkers were discussed by referring to the literature and databases.Conclusion The work may provide valuable clues for understanding the mechanisms of hypoxic preconditioning as well as showing the powerful combination of the LC-ESI-MS technique and the pattern recognition method for analysis of the biochemical changes of pathophysiologic conditions.