Advancements in computing and communication technologies, coupled with the ubiquitous availability of low‐cost embedded devices, have enabled the vision of cyber‐physical systems (CPSs). With the advent of advanced persistent threats, CPSs are more vulnerable to sophisticated cyber‐attacks that cause catastrophic damages, thereby necessitating the development of novel defence architectures for CPS security. This work presents an analytical framework based on noncooperative game theory to evaluate the trustworthiness of individual nodes that constitute CPSs. The proposed approach uses the Nash equilibrium solution to derive a minimum trust threshold score for the CPS nodes. The game‐theoretic framework is evaluated on a supervisory control and data acquisition system prototype to model the evolution of the trust scores its sensors. Furthermore, we apply the model on a simulated unmanned aerial vehicle (UAV) system and derive the trust threshold. The trust threshold represents the minimum trust score required to be maintained by individual UAV nodes. Nodes with trust scores below the threshold are potentially malicious and may be removed or isolated to ensure the secure operation of the system. The proposed approach is successfully implemented to detect malicious behavior in a simulated multiloop UAV control system with a fewer number of false‐positives, achieving a maximum accuracy of 98.85% across different scenarios.