More and more organizations turn to the construction of process models to support strategical and operational tasks. At the same time, reports indicate quality issues for a considerable part of these models, caused by modeling errors. Therefore, the research described in this paper investigates the development of a practical method to determine and train an optimal process modeling strategy that aims to decrease the number of cognitive errors made during modeling. Such cognitive errors originate in inadequate cognitive processing caused by the inherent complexity of constructing process models. The method helps modelers to derive their personal cognitive profile and the related optimal cognitive strategy that minimizes these cognitive failures. The contribution of the research consists of the conceptual method and an automated modeling strategy selection and training instrument. These two artefacts are positively evaluated by a laboratory experiment covering multiple modeling sessions and involving a total of 149 master students at Ghent University.