A population pharmacokinetic study of cyclosporine (CsA) was performed in liver transplant recipients. A total of 3731 retrospective drug monitoring data points at predose (C0) and 2 hours postdose (C2) were collected from 124 liver transplant recipients receiving CsA microemulsion. Population pharmacokinetic analysis was performed using the program NONMEM (nonlinear mixed-effect modeling). Various covariates potentially related to CsA pharmacokinetics were explored, and the final model was validated by a bootstrap method and by assessing the predictive performance using empiric Bayesian estimates. A one-compartment model with first-order absorption was considered. Population parameters of apparent clearance (CL/F) and volume of distribution were estimated as 23.1 L/h and 105 L, respectively. CL/F was influenced by four covariates: duration of CsA therapy (DT), hematocrit (HCT), and concurrent prednisone dose (PR). The final model for CL/F was fitted as follows: CL/F = 23.1 + 0.5 × (DT/200) - 0.07 × HCT + 0.04 × PR. The interindividual variability in CL/F, volume of distribution, and Ka calculated as coefficient of variation were 15.1%, 9.3%, and 66.0%, respectively. The intraindividual variability was 18.6%. The model fitted well with the observed data, and the bootstrap method guaranteed robustness of the population pharmacokinetic study model. Model validation was performed by a visual predictive check. Moreover, simulation was conducted to facilitate the individualized treatment based on patient information and the final model. The model to characterize population pharmacokinetic study of CsA provided better clinical individualization of CsA dosing in liver transplant recipients based on patient information and to assess patients' suitability for CsA therapy.