<p>Self-defense can be understood as the instinctive response of individuals in the face of cyber epidemic attacks. Nevertheless, research on individual-level preventive defense strategies and their impact on the network security situation from a dynamic perspective is notably lacking. This paper aims to construct a theoretical framework that depicts the interplay between self-defense strategies and cyber epidemic attacks. Firstly, we propose an individual-based model for cyber epidemic attacks, which incorporates the individual’s self-defense state. An analysis is then conducted to obtain the thresholds for both exponential and asymptotic stability within the cyber dynamical system. Moreover, taking advantage of the combination of cyber epidemic attacks and Markov decision processes, a framework is posited to investigate the optimal self-defense approach via the computation of individual expected utility. Ultimately, through the development of a simulation system that virtualizes cyber epidemic attacks and individual defense decisions, the simulation results corroborate the analytical conclusions pertaining to security thresholds and the optimal preventive self-defense strategy. These findings illuminate the significance of preventive self-defense in mitigating cyber epidemic attacks.</p>