Objective: The objective of this study is to develop population pharmacokinetics model of vancomycin in treatment of septicemia in infants under one year old, to investigate factors which have significant influences on pharmacokinetics of vancomycin, to optimize vancomycin dosing regimen, thereby to reduce the occurrence of adverse drug reaction of vancomycin, to improve the vancomycin therapy effect.Method: Non-linear mixed effect model (NONMEM) was employed to develop population pharmacokinetic model of vancomycin.The stability and predictive ability of final model were evaluated by normalized prediction distribution errors (NPDE),bootstrap method (Bootstrap) and external data.Final model was subjected to Monte Carlo simulation to determine the optimal dose under which drug concentration of a certain patient could be controlled within reference concentration.Result: A total of 205 trough and peak concentrations from 94 infants under 1 year old were analyzed.Exponential model was used to describe inter-individual variability of pharmacokinetic parameter.Residual error was better described by exponential model than mixed exponential and addition model.SCR and body weight were the major factors which impacted the PK parametric of vancomycin.The clearance would be larger when incorporate with Ceftriaxone.In this study, more than two model validation methods showed the good stability of the final internal model, with superior predictive performance, which can develop individualized dosing regimens for clinical reference.There is a greater probability that the trough concentrations would be lower than 5 mg/L when 3-9 months infants were given 10mg / kg q6h, so we should increase the dose.Conclusion: The established vancomycin PPK model of septicemia infants under 1 year old in this study was stable and effective, and has good predictive ability, which can be used for establishing clinical individualized dosing.