Air pollutants such as SO₂, NO_x, particulate matter, are emitted into the atmosphere and could contribute to acid rain production, smog and ground-level ozone. NO_x emissions have been decreasing in Portugal in road transport, industrial combustion and power stations. A better monitoring system and increasing the use of low carbon fuels integrated with renewable, affordable and clean energy could contribute to achieve at least 55% less net emissions in 2030 in comparison with 1990 levels according to the “Fit for 55” legislative package. In this work, an optimized machine learning random forest algorithm was developed and provided a good prediction of NO_x emissions from a natural gas combined cycle turbine power plant with 0.89% accuracy and a RMSE of 3.81. Furthermore, it was obtained a prediction of CO emissions with an accuracy of 0.79 and a RMSE of 1.08. The optimized machine learning random forest algorithm developed is a solution to predict pollutant emissions and contribute to mitigate and monitor exhaust gases emissions. Furthermore, it was concluded that ambient air temperature and turbine inlet temperature are essential to determine the production of NO_x, where lower inlet air temperature and gas turbine inlet temperature could mitigate NO_x emissions.