Self-Optimizing mechatronic systems which are able to react autonomously and flexibly to changing environments are one promising approach for the next generation of mechanical engineering systems. To render designing such systems possible, an approach is required which goes far beyond what is offered by today̢۪s standard tools for mechatronic systems. In this paper, we outline how a smooth integration between mechanical and software engineering methods and tools supports the design of verifiable, complex, reconfigurable mechatronic systems. The focus of the paper is on enabling the design and simulation of safe reconfigurable mechatronic systems, as reconfiguration is a critical prerequisite for self-optimization.