Process modelling has proved to be a powerful instrument to describe and manage the increasingly complex processes within and across enterprises. Yet, it requires a significant amount of manual work to create, adapt and maintain process models. This workload could be reduced if the process models are automatically generated and adapted. Semantic Business Process Management in combination with planning approaches can contribute to the solution. In this paper we describe that existing planning algorithms show drawbacks for this application and therefore introduce an innovative algorithm that is suitable for the semanticbased planning of process models.