Purpose Population pharmacokinetics (PK) of sepantronium bromide (YM155) was characterized in patients with non-small cell lung cancer, hormone refractory prostate cancer, or unresectable stage III or IV melanoma and enrolled in one of three phase 2 studies conducted in Europe or the U.S. Method Sepantronium was administered as a continuous intravenous infusion (CIVI) at 4.8 mg/m(2)/day over 7 days every 21 days. Population PK analysis was performed using a linear one-compartment model involving total body clearance (CL) and volume of distribution with an inter-individual random effect on CL and a proportional residual errors to describe 578 plasma sepantronium concentrations obtained from a total of 96 patients by NONMEM Version VI. The first-order conditional estimation method with interaction was applied. Results The one-compartment model with one random effect on CL and two different proportional error models provided an adequate description of the data. Creatinine clearance (CLCR), cancer type, and alanine aminotransferase (ALT) were recognized as significant covariates of CL. CLCR was the most influential covariate on sepantronium exposure and predicted to contribute to a 25 % decrease in CL for patients with moderately impaired renal function (CLCR = 40 mL/min) compared to patients with normal CLCR. Cancer type and ALT had a smaller but nonetheless significant contribution. Other patient characteristics such as age, gender, and race were not considered as significant covariates of CL. Conclusions The results provide the important information for optimizing the therapeutic efficacy and minimizing the toxicity for sepantronium in cancer therapy.