In a market environment increasingly characterized by oversupply and intensified product homogenization， manufacturing firms urgently need to enhance product adaptability to quickly respond to dynamic changes in user demands and strengthen their market competitiveness. It took Midea Group as a case study to explore the mechanism of AI-driven scenario-based product adaptive innovation. The findings reveal that the AI-driven scenario-based product adaptive innovation process consists of three stages： scenario-based functionality identification， functionality matching， and functionality adjustment. Scenario-based dynamic capability serves as the foundational capability for innovation， encompassing three dimensions of scenario perception capability， scenario capture capability， and scenario reconfiguration capability. AI drives product adaptive innovation through correlation analysis， intelligent insights， and continuous learning. It constructs a theoretical framework of AI-driven scenario-based product adaptive innovation， proposes scenario-based dynamic capability as its foundational capability， and clarifies AI’s role in this process. By doing so， it advances research in product innovation， extends the application of dynamic capability theory， and deepens the understanding of AI’s enabling role in product innovation， providing theoretical insights and practical implications for manufacturing firms in the era of intelligent digitalization.