This study compares ARX and ARMAX models for forecasting COVID-19 case dynamics. Both models have been frequently used in time series forecasting and are considered here for their efficacy to estimate the spread of COVID-19. The ARX model utilizes delayed values of the dependent variable as well as exogenous inputs, whereas the ARMAX model includes a moving average component to account for random disturbances and noise in the data. This study focuses on each model’s performance, accuracy, and stability, emphasizing how the presence of external factors like movement patterns, government actions, and vaccination rates influences predictive ability. The work is based on real-world COVID-19 data, and the models are evaluated using important metrics such as root mean square error (RMSE) and goodness of fit. The results show that, while both models are successful, the ARMAX model makes more accurate predictions in the face of noise and changing external factors. This comparison emphasizes the benefits and drawbacks of each method and provides insights on developing time series models for pandemic forecasting.