Survival curves of Enterobacter sakazakii inactivated by high hydrostatic pressure were obtained at four pressure levels (250, 300, 350, and 400 MPa), at temperatures below 30 degrees C, in buffered peptone water (BPW; 0.3%, wt/vol) and infant formula milk (IFM; 16%, wt/vol). A linear model and four nonlinear models (Weibull, log-logistic, modified Gompertz, and Baranyi) were fitted to the data, and the performances of the models were compared. The linear regression model for the survival curves in BPW and IFM at 250 MPa has fitted regression coefficient (R2) values of 0.940 to 0.700, respectively, and root mean square errors (RMSEs) of 0.770 to 0.370. For the other pressure levels, the linear regression function was not appropriate, as there was a strong curvature in the plotted data. The nonlinear regression models with the log-logistic and modified Gompertz equations had R2 values of 0.960 to 0.992 and RMSE values of 0.020 to 0.130 within pressure levels of 250 to 400 MPa, respectively. These results indicate that they are both better models for describing the pressure inactivation kinetics of E. sakazakii in IFM and BPW than the Weibull distribution function, which has an R2 minimum value of 0.832 and an RMSE maximum value of 0.650 at 400 MPa. On the other hand, the parameters for the Weibull distribution function, log-logistic model, and modified Gompertz equation did not have a clear dependence on pressure. The Baranyi model was also analyzed, and it was concluded that this model provided a reasonably good fit and could be used to develop predictions of survival data at pressures other than the experimental pressure levels in the range studied. The results provide accurate predictions of survival curves at different pressure levels and will be beneficial to the food industry in selecting optimum combinations of pressure and time to obtain desired target levels of E. sakazakii inactivation in IFM.