Interdisciplinary cooperation willingness serves as the core micro-foundation supporting the resilience of technological innovation, yet its practical dilemma of "high initial willingness but low sustainability" restricts the improvement of China's innovation system efficiency. To explore this dynamic fluctuation law, based on evolutionary game theory and interdisciplinary collaboration theory, this paper focuses on three core mechanisms: academic power balance, benefit distribution, and expected returns, constructs a tripartite evolutionary game model between research platforms and researchers from different disciplines, and analyzes the impact of key variables on cooperation willingness through simulation. The study finds that: Academic power balance is the core of stable cooperation; moderate disciplinary differences can stimulate cooperation willingness, while power imbalance prolongs the stabilization cycle, with task centrality serving as a regulatory tool. The fairness of benefit distribution determines cooperation resilience; unfair distribution is prone to cause systemic risks, which can be adjusted by subdividing implicit and explicit benefits in line with the cooperation stages. Appropriate punishment mechanisms are more conducive to guiding cooperation than incentives. Corresponding policy suggestions such as hierarchical response, power checks and balances, fair distribution, and scientific rewards and punishments are put forward, providing a theoretical basis for building the stability of interdisciplinary cooperation and a decision-making reference for improving the more resilient technological innovation system.