Under the distributed adaptive discrete event-triggered communication scheme (D-ADETCS), this study investigates the co-design problem between multimodal integrated security control and communication network for a class of industrial cyber–physical systems (ICPS) subject to both multi-source stealthy false data injection (FDI) attacks and actuator faults, from both offensive and defensive perspectives. First, by analysing the correlation between the variation rate of the event-triggering threshold and the system operating state, a novel adaptive discrete event-triggered threshold function is designed. An integrated security control architecture driven by the D-ADETCS is then constructed by deploying the ADETCS at both ends of the system network. Second, considering the characteristics of FDI attacks such as stealthy, effectiveness, and target selectivity, a multi-source stealthy FDI attack model supporting collaboration among multiple network nodes is established. Based on this model, a distributed data repairer is designed using a lightweight gradient boosting machine optimized by the Greater cane rat algorithm to counter multi-source covert FDI attacks. Third, using methods such as the Lyapunov-Krasovskii functional, the proofs for a robust observer and a multimodal integrated security controller are derived. Subsequently, control modes are partitioned using the fast density peaks clustering algorithm, and flexible switching among control modes is achieved through a fuzzy control strategy. Finally, simulation results demonstrate that, compared with existing methods, the proposed approach exhibits significant advantages in terms of integrated security control and optimized communication resource allocation.