Facing the growing complexity of car electronics, carmakers and automotive electronic suppliers are looking for efficient methods to develop, verify and validate electronic modules. In fact, they focus on the software part of these modules since it accounts for more than 80% of the total number of problems detected on these modules. In this context, we achieved our research project with the aim of proposing a global approach able to improve the quality of automotive embedded software. We started with an audit of the software practices currently used in automotive industry and we pinpointed potential levers to improve the global software quality. Based on the results of the audit and the literature review related to software quality, we developed a new Model-Based Testing approach in order to automatically generate test cases for a software product. This approach takes into account most of the automotive software constraints and context. The results of our experiments reveal significant improvement in software quality: more bugs are detected earlier and in less time.