Traditional cyber defense strategies rely on a linear approach that involves detecting threats, selecting defenses, and mitigating attacks; yet, they struggle with emerging, unrecognized, and advanced threats. In search of a more robust solution, researchers have explored innovative strategies to maintain cybersecurity in a network without prior knowledge of the adversary or the specific attack being executed. One such strategy is known as Moving Target Defense (MTD). Leveraging Bayesian Stackelberg game theory, we establish optimal strategies for the defender and adversary, showcasing how the defender can reduce costs by steering attacks away from higher criticality nodes. This approach helps the defender implement a novel MTD logic model for either diversion or minimization of the attack damages. We use simulation results to show how our approach surpasses previous strategies. Our approach offers improvements in managing a multitude of resources. The new approach, while not addressing the known drawbacks, lays the foundation for more advanced MTD models that can incorporate a more detailed representation of system resources.