In this paper, an optimization procedure for complex manufacturing processes is presented. The procedure is based on advanced empirical modeling techniques and will be presented in two parts. The first part comprises the selection and generation of the empirical surrogate models. The process organization and the design of experiments are taken into account. In order to analyze and optimize the processes based on the empirical models, advanced methods and tools are presented in the second part. These tools include visualization methods and a sensitivity and robustness analysis. Moreover, the obtained surrogate models are used for a model-based multi-objective optimization in order to explore the gradation potential of the processes. The procedure is applied to two thermo-mechanical processes for the functional gradation of polymers - a monoxiale stretching of polycarbonate films and a compression moulding process for polypropylene sheets.