Oxycodone has been one of the most common opioids found in drug-induced poisoning deaths in Swedensince 2020. A better understanding of oxycodone’s metabolism could allow us to develop new tools thatcould prevent oxycodone related deaths. One way to improve understanding of oxycodone metabolism isto develop a mathematical model of its pharmacokinetics. A pharmacokinetic model describes the concentrations of oxycodone and its metabolites in a person’s blood or urine over time. The concentration ofoxycodone and its metabolites can provide information if, for example, a person acquires dangerous concentrations of oxycodone in their body. What is troubling is that current pharmacokinetic models of oxycodone metabolism have shortcomings in describing all relevant metabolites, accounting for the CYP2D6phenotype, and incorporating individual anthropometric measurements. To address these problems, a mathematical model of oxycodone metabolism was developed by using iterative model-based hypothesis testing. The developed model could describe the three main metabolic pathways: O-demethylation,N-demethylation, and 6-keto-reduction, over 24 hours for oxycodone and its metabolites in the blood.The model could do this with respect to anthropometric measurements and the CYP2D6 phenotype.During model development and implementation of CYP2D6 phenotype consideration, model evaluationrevealed that CYP2D6 metabolisers exhibited overall poorer oxycodone metabolism not only in the Odemethylation pathway but also in the reduction and N-demethylation pathways. This behaviour wascontrary to what the literature suggested about how the poor metabolisers should behave.Continuing, when attempting to describe both blood and urine dynamics simultaneously, the modelfailed to explain the urinary dynamics. However, because the model accurately described oxycodonedynamics in the blood, it could potentially be used by forensics in the future to predict the time ofintake, dose size, and route of administration from a single sample collected from living patients. Beingable to determine the time of intake, dose size, and route of administration could be useful in futureforensic cases where such information is scarce. The model could also be used in the future to createpatient-specific dosing plans to prevent accidental overdosing cases.