This study utilises a SIR model variant to analyse the dynamics of COVID-19 transmission in Brazil. The model incorporates compartments for different population states and optimises infection and loss of immunity rates to achieve a closer fit to real-world data, improving the characterisation of diseases’ behaviours and informing public health strategies. Data from the Our World in Data portal, including information on vaccines, variants’ dominance, and new developed cases, is used to calibrate the model by minimising the quadratic error between its projections and actual case numbers. Key parameters include infection rates, loss of immunity, vaccination, and the duration of infection and exposure. The study addresses mathematical challenges and discusses non-mathematical variables that influence outcomes. The goal is to explain infection dynamics in Brazil over time by determining infection and loss of immunity rates through simulations, considering various scenarios, including the impact of vaccination.