Oxycodone is an opioid analgesic with several pharmacologically active metabolites and relatively narrow therapeutic index. Cytochrome P450 (CYP) 3A4 and CYP2D6 play major roles in the metabolism of oxycodone and its metabolites. Thus, inhibition and induction of these enzymes may result in substantial changes in the exposure of both oxycodone and its metabolites. In this study, a physiologically based pharmacokinetic (PBPK) model was built using GastroPlus™ software for oxycodone, two primary metabolites (noroxycodone, oxymorphone) and one secondary metabolite (noroxymorphone). The model was built based on literature and in house in vitro and in silico data. The model was refined and verified against literature clinical data after oxycodone administration in the absence of drug–drug interactions (DDI). The model was further challenged with simulations of oxycodone DDI with CYP3A4 inhibitors ketoconazole and itraconazole, CYP3A4 inducer rifampicin and CYP2D6 inhibitor quinidine. The magnitude of DDI (AUC ratio) was predicted within 1.5‐fold error for oxycodone, within 1.8‐fold and 1.3–4.5‐fold error for the primary metabolites noroxycodone and oxymorphone, respectively, and within 1.4–4.5‐fold error for the secondary metabolite noroxymorphone, when compared to the mean observed AUC ratios. This work demonstrated the capability of PBPK model to simulate DDI of the administered compounds and the formed metabolites of both DDI victim and perpetrator. However, the predictions for the formed metabolites tend to be associated with higher uncertainty than the predictions for the administered compound. The oxycodone model provides a tool for forecasting oxycodone DDI with other CYP3A4 and CYP2D6 DDI perpetrators that may be co‐administered with oxycodone.