From the perspective of the interactive cooperation among subjects, this paper portrays the process of cooperative innovation in industrial cluster, in order to capture the correlated equilibrium relationship among them. Through the utilization of two key tools, evolutionary stable strategy and replicator dynamics equations, this paper considers the cost and gains of cooperative innovation and the amount of government support as well as other factors to build and analyze a classic evolutionary game model. On this basis, the subject’s own adaptability is introduced, which is regarded as the system noise in the stochastic evolutionary game model so as to analyze the impact of adaptability on the game strategy selection. The results show that, in the first place, without considering subjects’ adaptability, their cooperation in industrial clusters depends on the cost and gains of innovative cooperation, the amount of government support, and some conditions that can promote cooperation, namely, game steady state. In the second place after the introduction of subjects’ adaptability, it will affect both game theory selection process and time, which means that the process becomes more complex, presents the nonlinear characteristics, and helps them to make faster decisions in their favor, but the final steady state remains unchanged.