A variety of quality criteria play an important role in the production of powder coatings. In addition to mechanical properties such as abrasion, scratch and chemical resistance, optical properties such as degree of gloss, color and surface roughness are among the most important quality parameters. These factors are influenced by countless parameters such as type of compounder, screw configuration, reactivity of the resin as well as process parameters. Processing is very complex due to the high reactivity of the selected polymer systems. Therefore optimal process design as well as scale-up becomes more difficult. Influence of different plant and process parameters on various optical quality criteria such as color, gloss and surface roughness as well as mechanical properties have been investigated for different types of powder coatings. For this purpose, tests were carried out on co-rotating twin screw extruders of different types and sizes on laboratory as well as on production scale. In addition the influence of the screw geometry was also investigated by varying the kneading and mixing elements. Furthermore, the changes in the reactivity by means of DSC measurements of the different powder coatings were also determined. The measured data were analyzed using Smart Data Mining and Heuristic optimization tools based on evolutionary algorithms. Therefore the development of prediction models as well as the evolution of decision trees for classification of the determined data was possible. Symbolic regression based on genetic programming allowed the development of analytical equations to describe selected product properties taking into account the complex relationships. Additionally numerical calculations for the description of the process were carried out in order to describe the characteristic operating behavior of co-rotating twin screw extruders during powder coating production and thus enable the scale-up of process conditions from laboratory to production scale.