Caspofungin is an echinocandin antifungal agent licensed as a first-line therapy for invasive candidiasis in patients with moderate to severe illness or recent exposure to azoles. In this study we developed a whole-body physiology-based pharmacokinetics (WB-PBPK) model to predict the pharmacokinetics (PK) of caspofungin, and combined with Monte Carlo simulation (MCS) to optimize clinical dosage regimens of caspofungin in different kinds of patients. A WB-PBPK model of caspofungin was built and validated with raw data from 4 previous trials of general patients, intensive care unit (ICU) patients with Child-Pugh B, ICU patients on continuous renal replacement therapy, mild and moderate hepatic insuffciency (HI) patients. MCS was used to optimize clinical dosage regimens of caspofungin in these patients. A cumulative fraction of response (CFR) value of ≥90% was considered to be the minimum for achieving optimal empirical therapy. The simulated results of the WB-PBPK model were in good agreement with observed values of all trials. For general and ICU patients with caspofungin 70/50 mg, AUC and Cmax were decreased with the increase of body weight (BW) and showed great variation. MCS showed all general patients achieved CFR≥90% regardless of BW. But not all ICU patients with higher BW (≥70 kg) could achieve CFR≥90%. Compared with standard dosage regimens in general patients, caspofungin 70/35 mg in ICU patients with Child-Pugh B achieved significantly decreased AUC and Cmax, but obtained similar AUC and Cmax in moderate HI patients with Child-Pugh B. The WB-PBPK model of caspofungin is able to predict PK of all populations correctly. The combined WB-PBPK model with MCS can successfully optimize clinical dosage regimens of caspofungin in all patient populations.