This work proposes a method for designing a model based on the Multi-Agent System (MAS) and Digital Twin (DT) concepts to study the cyber-physical systems security. When Cyber-Physical Systems (CPS) are used in a network to address a complex problem, they present a uniquely wide vulnerability challenges as their attack surface ranges from hardware and physical attacks to software attacks and even including network attacks. To meet these challenges, we explored several approaches to combine MAS and DT with the aim to benefit from the scalability and adaptability of MASs and the enhanced modelling of DTs. To showcase our approach, we present its application to detect blackhole attacks (a kind of attack in which one or more nodes attract all communications and not forward them, mimicking an error in the network) in a simulated smart home environment. As results are promising, we conclude and discuss future research perspectives in allying DTs and MASs for managing the security of CPSs. The code of experimentations is available on Ghithub. https://github.com/ArthurBadet/PAAMS-Code