Aiming at the harm of flexible and changeable network malicious code to network security, it is proposed to combine honeypot entrapment with the characteristics of IoT interconnectivity and high information real-time, and introduce deception nodes on the basis of the classical SIR propagation model, which is able to utilize honeypots to entrap the infected nodes to generate deception information, and then disseminate the deception information through IoT to achieve the immunity of the nodes. Then, the dynamics analysis and numerical simulation and simulation of the model are carried out, and the experimental results show that the total number of infected nodes at the peak point of this model decreases by 32.35% compared with the classical SIR model when the propagation threshold is less than 1 and the trapping rate takes the value of 0.008 and the feedback rate takes the value of 0.0000003; and the total number of infected nodes at the peak point in this model decreases by 12.04% when compared to the existing SIDR model that takes into account the honeypot trapping technique.