Crucial for training automotive mechatronics technicians (AMTs) is enabling them to diagnose car malfunctions. AMTs are particularly successful when they base their diagnostic process on a mental model of the affected automotive system. Still, only few AMT apprentices master such diagnoses after their apprenticeship. Therefore, we created a simulation-based learning environment with modeling examples to teach AMT apprentices a diagnostic strategy that builds on mental models. Following design-based research guidelines, we formatively evaluated our learning and testing materials by expert judgments and a small study during the development of the learning environment. Finally, an evaluation study showed that the learning environment promoted apprentices' knowledge about the diagnostic strategy. However, they could not transfer their knowledge to diagnostic problem-solving. Overall, the apprentices evaluated the learning environment positively, except it was considered too long and repetitive. Reasons for the outcomes as well as possible further developments of the learning environment are discussed.