Mechatronic systems today have to fulfill increasingly complex tasks in diverse and often safety-critical situations. In order to cope with this complexity, the design of the system is typically based on scenarios, in which the engineers describe which sequences of events may, must, or must not happen in certain situations. Scenarios allow humans to conceive complex requirements. However, it may happen that contradictions are introduced among the scenarios, and thus the specification becomes inconsistent. If such inconsistencies are not detected early, this may require costly iterations in the systems development or it may lead to flaws in the system.In the scope of this thesis, a method was developed for finding inconsistencies in scenario-based specifications of mechatronic systems. Modal Sequence Diagrams (MSDs) were extended so that now real-time requirements and environment assumptions can be formulated. A technique was developed that maps the problem of finding inconsistencies in such MSD specifications to the problem of synthesizing winning strategies in two-player games. This way, an existing, efficient algorithm can be employed for consistency analysis. Moreover, this thesis presents a formal technique for decomposing the synthesis problem.Furthermore, in order to find inconsistencies in specifications of dynamic systems, this thesis presents concepts for improving the play-out algorithm. This algorithm allows for the simulation of MSD specifications, but it may run into avoidable violations of the specification. By combining the play-out with strategies that could be successfully synthesized from parts of the specification, the simulation produces less avoidable violations, thus giving the engineers more reasons to suspect inconsistencies if violations occur.