As one of the most commonly used modes of transportation, the majority of automobiles are in the L2 to L3 stage of humanmachine codriving technology development. Before the emergence of L5level autonomous driving technology, “humanmachine codriving” remains the dominant driving method, with its various invehicle systems and interaction methods continuously being improved. The integration of multimodal interaction with automotive technologies is bound to ignite new "sparks" as a future design trend. This paper firstly summarizes the research on multimodal interaction design for invehicle systems, covering directions such as fatigue warning, collision warning, lane departure warning, intelligent takeover reminder, and intelligent parking. Subsequently, it analyzes the natural interaction methods of invehicle AI multimodal interaction design, including multiscreen interaction, touch interaction, gesture interaction, voice interaction, facial expression interaction and eye movement interaction. The article uses literature research and case studies to explore how to improve the driver's comfort in the context of safety and emotional aspects, and anticipates the applications and future trends of multimodal interaction design for invehicle systems. Finally, it is concluded that the integration of appropriate and effective interaction methods will improve the safety and driving comfort of various invehicle systems and applications. The introduction of multimodal interaction is destined to become a trend in automotive development.