OWA (Ordered Weighted Averaging) operator provides a general class of parameterized mean type aggregation operators. It is an important tool, having ability to reflect the strategy of a Decision Maker in decision making problems. The idea of using a stress function to generate the OWA weights was introduced in this study. We introduce a new fuzzy c-regression models (FCRM) based approach to obtain stress function for generalization of OWA weights. This generalization was realized by using FCRM as learning method to a stress function which is related the known OWA weights.