We developed a pharmacokinetic/pharmacodynamic (PK/PD) mathematical model that fits voriconazole time-kill data against Candida isolates in vitro and used the model to simulate the expected kill curves for typical intravenous and oral dosing regimens. A series of Emax mathematical models were used to fit time-kill data for two isolates each of Candida albicans, Candida glabrata and Candida parapsilosis. PK parameters extracted from human data sets were used in the model to simulate kill curves for each isolate. Time-kill data were best fit by using an adapted sigmoidal Emax model that corrected for delays in candidal growth and the onset of voriconazole activity, saturation of the number of Candida and the steepness of the concentration-response curve. The rates of maximal killing by voriconazole (kmax) were highly correlated with the growth rates (ks) of the isolates (Pearson's correlation coefficient=0.9861). Simulations using PK parameters derived from the human data sets showed fungistatic effects against each of the isolates. In conclusion, we demonstrated that the activity of voriconazole against Candida isolates can be accurately described using a mathematical model. In the future, it might be possible to devise optimal dosing regimens of voriconazole using the model and PK data collected in vivo.