<p>This paper seeks to solve the issue of asynchronous security controller design for a class of nonlinear stochastic cyber physical systems (CPSs) using a probabilistic dynamic event-triggered strategy to deal with actuator failures and parameter dependence on multiple cyber-attacks. The probabilistic-dynamic-memory event-triggered protocol (PDMETP) is a unique event-triggered system that is proposed for the first time. Meanwhile, T-S fuzzy singular semi-Markov jump systems are utilized to characterize nonlinear stochastic CPS. Furthermore, to address the asynchronous difficulties, we use the Hidden-Markov-Model to express the asynchronous state between the security controller and the original system. Furthermore, it is assumed that the conditional probability is only partially known. The physical layer of actuator defects is added to imitate the real scenario of the process of production. Then, using the random Lyapunov functional and the free weighting matrix approach, the closed-loop system’s regular, impulse free, and stochastically stable conditions are determined. Finally, two numerical and practical examples demonstrate the method’s effectiveness and application.</p>