The paper presents a framework for model-based testing of automotive distributed system and a method of automatic assignment of testing priorities used within the framework. The proposed method utilizes classifiers for automatic assignment of testing priorities to specific parts of the tested system. The paper also introduces a set of extraneous data accompanying the modeling language that are exploited by the proposed method during the classification process. It is shown, that advantages of the presented approach, such as lower requirements for the testing operators' knowledge, are valuable for the automotive distributed systems testing process.