In today's industry, machine learning (ML) and additive manufacturing (AM) are revolutionary technologies. By modelling surface roughness based on thermal analysis and predicting its surface roughness value using machine learning, this work attempts to improve the surface quality of 3D printed objects. The optimization of key parameters such as layer height (LH), printing speed (PS), nozzle temperature (NT), and infill density (ID) will take place.ML algorithms such random forest regressor, XG-Boost, support vector machines, and linear regression can be used to make the prediction. The PLA+ material characterization will also be looked at.To analyze parameter effects, experiments employ Taguchi's Design of Experiment with orthogonal array, and machine learning methods will be used to determine which model is the most correct.The work focusses on LH, ID, PS, NT, and platform temperature as the five input parameters that affect layer geometries. By optimizing AM processes, advanced machine learning algorithms seek to improve the surface quality of 3D printed items.