Objective This study was designed to validate the utility of a population pharmacokinetic model established for vancomycin in patients with severe neurosurgical disease . Methods The clinical data including patient gender , age , body weight ,serum creatinine and albumin were collected retrospectively from patients in Nanjing Drum Tower Hospital to calculate the steady trough concentration of vancomycin using the previously established pharmacokinetic model .The predicted value was compared with the actual value .Results During the period from March 2013 to March 2014 ,53 blood samples with serum trough concentration of vancomycin were collected from 42 patients .The average trough concentration of vancomycin was 10 .9 mg/L (range from 1 .6 to 49 .1 mg/L) .The predicted trough level of vancomycin based on the population pharmacokinetic model was significantly correlated to the actual value(r=0 .857 ,P<0 .001) .The mean absolute percentage error was 0 .407 9 . The confidence interval was 9 .36‐14 .07 for the predicted values ,and 8 .92‐14 .32 for the actual values .Conclusions The pharmacokinetic model is valid and useful for planning intravenous dose of vancomycin in patients with severe neurosurgical disease .Large error (about 30% ) was observed in estimation of body weight due to coma .Reduced renal function following contrast agent and/or diuretic drug has an impact on the predicted results . The accuracy of prediction can be increased to nearly 70% after adjusting the covariates .