This thesis investigates the prediction of road travel speed in Gothenburg, Sweden, using regression analysis. Leveraging a comprehensive dataset from the Swedish Transport Administration, the study integrates temporal variables, route information, and weather conditions to develop a prediction model. The XGBoost regressor, known for its effectiveness with complex datasets, was employed to forecast the speed for upcoming road segments. Evaluation metrics, including Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R-squared (R²), indicated high predictive accuracy and robustness. Despite the model's high performance, limitations such as the absence of data on road accidents and maintenance activities, as well as a restricted geographical focus, were noted. The findings offer valuable insights for enhancing transportation planning, route optimization, and traffic management in urban settings.