Central bank management of money creation and credit steering is fundamental to financial stability and the dynamism of macroeconomic variables. While existing literature has extensively examined bank–firm relationships or general financial regulation, researchers have paid less attention to the direct, repeated strategic interaction between the central bank and heterogeneous commercial banks. This study employs an evolutionary game theory framework to analyze strategic interactions in the credit creation process between a central bank and two distinct bank populations: large state-owned banks and small private banks. Unlike static or fully rational models, this framework incorporates bounded rationality and replicator dynamics, where successful strategies spread through learning and imitation. This study explicitly models key institutional and behavioral parameters—including supervision intensity, reputational cost, learning rate, and penalty for deviation. The results demonstrate that increasing supervision intensity and reputational costs significantly reinforce disciplined, trust-based behavior within the banking system, thereby enhancing financial stability and overall macroeconomic efficiency. Conversely, weak supervision and low reputational costs lead to an undesirable equilibrium of pervasive opportunism and credit contraction. These findings provide a robust theoretical foundation for designing more effective, behaviorally informed credit management policies in mixed banking systems.