The COVID-19 pandemic has highlighted the critical need for accurate disease forecasting models to guide public health interventions and policy decisions. This thesis investigates the application of the Autoregressive Integrated Moving Average (ARIMA) model for forecasting the spread of COVID-19. The ARIMA model, known for its robustness in time series analysis, is utilized to predict daily confirmed cases, recoveries, and deaths across various geographical regions. The study begins with a comprehensive data preprocessing phase, addressing issues such as missing values, outliers, and the need for stationary time series data. Subsequently, we fit the ARIMA model to historical COVID-19 data, optimizing its parameters using techniques such as grid search and cross-validation to ensure the best predictive performance. Our findings demonstrate that the ARIMA model can effectively capture the temporal dynamics of COVID-19 spread, offering reliable short-term forecasts. The model's performance is evaluated using standard metrics such as Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE), and results are compared against other forecasting methods including exponential smoothing and machine learning-based approaches.