In the knowledge economy era, both academia and industry alike have been focused on the impact of market knowledge on innovation performance. However, research addressing the inconsistencies in empirical findings about its impact (such as a negative or insignificant affect) is scarce. Moreover, ambidexterity (exploratory learning/exploitative learning) is one of the central topics in knowledge management. This article marks the first endeavor to adopt the knowledge-based theory and the ambidextrous learning perspective, attempt to create a theoretical framework of knowledge-learning-innovation, and thoroughly examine related causal relationships between different dimensions of the constructs. The empirical results demonstrated the following: market knowledge depth directly and positively impacts process innovation and product innovation; market knowledge breadth indirectly and positively impacts process innovation and product innovation; and there is no significant difference in the effects of the two types of knowledge on the two types of innovation performance. Ambidextrous learning directly and positively affects process innovation and product innovation; ambidextrous organizational learning mediates the effect of market knowledge breadth on process innovation and product innovation, and this mediating effect is more pronounced with exploitative learning; ambidextrous organizational learning does not mediate the effect of market knowledge depth on both types of innovation performance.