With the proliferation of short-range networks and the prevalence of devices connected to these networks, the Internet of Things (IoT) has held the promise of improving our lives. Devices-to-device communication is considered to be an important part of the IoT. To address the limitation of physical layer security approaches based on single-antenna systems, cooperative jamming is a promising approach to enhance efficient cooperation among devices. In this paper, we study the cooperative interference problem of a two-layer network in which the upper social network (S*) is composed of people with different social ties and the lower layer network (P*) is composed of various physical devices. S* overlays P* by a connecting degree (Cd). We introduce game theory to simulate the cooperation willingness of jammers in S*, and prove the existence of Nash Equilibrium (NE) and design an algorithm to calculate the secrecy outage probability (SOP). Moreover, we introduce a susceptible-infective-recovery (SIR) spreading model to evaluate the performance, which believes that each jammer’s initial state is likely to participate rather than certainly participate in the cooperation. Each jammer is considered to have three states: probable participation (P-state), obligatory participation (O-state) and non-participation (N-state). Experiments show that when social ties are strong among devices, jammers would be more willing to cooperate, contribute more to communication quality and have lower computational complexity on two-lay net-work. Our methods make jammers have more choices for updating status, and show that results would be better than without SIR characteristics under the same number of jammers.