Pazopanib is a multi-targeted anticancer tyrosine kinase inhibitor. This study was conducted to develop a population pharmacokinetic (popPK) model describing the complex pharmacokinetics of pazopanib in cancer patients. Pharmacokinetic data were available from 96 patients from three clinical studies. A multi-compartment model including (i) a complex absorption profile, (ii) the potential non-linear dose-concentration relationship and (iii) the potential long-term decrease in exposure was developed. A two-compartment model best described pazopanib pharmacokinetics. The absorption phase was modelled by two first-order processes: 36 % (relative standard error RSE 34 %) of the administered dose was absorbed with a relatively fast rate (0.4 h RSE 31 %); after a lag time of 1.0 h (RSE 6 %), the remaining dose was absorbed at a slower rate (0.1 h RSE 28 %). The relative bioavailability (rF) at a dose of 200 mg was fixed to 1. With an increasing dose, the rF was strongly reduced, which was modelled with an E (maximum effect) model (E was fixed to 1, the dose at half of maximum effect was estimated as 480 mg RSE 23 %). Interestingly, the plasma exposure to pazopanib also decreased over time, modelled on rF with a maximum magnitude of 50 % (RSE 27 %) and a first-order decay constant of 0.15 day (RSE 43 %). The inter-patient and intra-patient variability on rF were estimated as 36 % (RSE 16 %) and 75 % (RSE 22 %), respectively. A popPK model for pazopanib was developed that illustrated the complex absorption process, the non-linear dose-concentration relationship, the high inter-patient and intra-patient variability, and the first-order decay of pazopanib concentration over time. The developed popPK model can be used in clinical practice to screen covariates and guide therapeutic drug monitoring.