In this chapter, we discussed how to model and analyze settings in which multiple decision-makers with misaligned interests can dynamically affect the spread of a malware in a large network. Specifically, we presented a framework based on differential game theory and introduced the notion of saddle-point strategies as mutually optimal dynamic controls in a game between a malware and a network defender. Through a stylized example, we showed how similar techniques from optimal control theory as in the previous chapter can help discover structural characteristics of such strategies in the absence of a closed-form solution. In particular, we showed that a killing worm facing a dynamic and strategic defense is still best off to have an initial phase of maximum intensity with no killing even while losing some of the infected nodes to be recovered by the system until a certain time after which it should start killing with maximum intensity. On the other hand, the defender should aggressively reduce the transmission range of the nodes and patch intensely right from the beginning until threshold times, after which it should relax the network operation to normal and subsequently stop patching.