Elderly users face a significant "digital divide" when interacting with automotive human-machine interfaces. The emergence of embodied intelligence offers innovative ideas for developing interfaces that better align with the natural cognition and interaction habits of elderly users. This study, from the theoretical perspective of embodied intelligence, systematically analyzes the driving characteristics of elderly users and their core interaction needs and pain points in in-vehicle scenarios through user research. Furthermore, it proposes a set of age-friendly design principles for automotive HMI based on embodied intelligence, providing a theoretical foundation and practical approach for creating inclusive, safe, and enjoyable next-generation automotive interaction experiences.