OBJECTIVES:  The NZ-MOA model is a population-based state-transition microsimulation model of knee osteoarthritis (OA) in New Zealand (NZ) currently under development. In this study, we aimed to validate the performance of the model by evaluating its ability to predict the incidence of total knee replacement (TKR) surgery in NZ. METHODS:  In the NZ-MOA model, individuals drawn from the New Zealand adult population are followed through an annual sequence of OA disease progression and associated health-related quality of life (HRQoL) changes. HRQoL is assessed using the six dimensions of the SF-6D instrument. Based on observed clinical criteria determining eligibility for publically-funded TKR, we defined SF-6D thresholds at which simulated individuals would receive TKR. Model performance was evaluated by comparing the predicted number of TKR with the number actually performed in two regions: the Southern District Health Board (SDHB) catchment area, a region covering a population of approximately 300,000 in the lower South Island of NZ, from which data on clinical criteria were obtained; and nationwide. RESULTS:  Applying observed SF-6D thresholds (limitations in Physical Functioning, Role Limitations due to physical health, and moderate or greater Pain, together with radiographic OA), the model predicted an annual rate of 191.1 knee replacements per 100,000 population aged 45+ in the SDHB region, compared with an observed rate of 196.5/100,000. Nationwide, slightly less stringent eligibility criteria are applied, resulting in a rate of 229.9/100,000; in the NZ-MOA model, this rate corresponded to a threshold in which approximately 50% of those with OA and meeting the pain criterion together with either of the physical limitations criteria also received TKR. CONCLUSIONS:  The NZ-MOA model was able to accurately predict the annual incidence of TKR. The model will be used in further work to evaluate the cost-effectiveness of various interventions, including TKR, for the treatment of OA.