A model for distributed computing called fog computing extends the cloud computing to the edge of network. By distributing computing power over several network edge devices, fog computing can reduce latency, boost data security and privacy that will enhance the overall system performance. A primary security concern is the creation of a trusted and reliable fog network. The process of creating and preserving trust between entities in a system, such as between users and service providers, is referred to as trust management. To handle distributed nature of fog environment, distributed and flexible trust management system is required. Reinforcement learning a branch of machine learning can be used to self-train each fog in network to identify malicious node and also make network more reliable. Machine learning (ML) can also be used to assess fog nodes’ reliability by examining their reputation, performance, and behavior. By using data from reliable fog nodes, machine learning algorithms can be trained to identify patterns of behavior that indicate reliability. These algorithms can then be used to evaluate the behavior of unknown fog nodes and assign them a trust score. This study outlines several unresolved challenges of machine learning-based security and trust issues in fog computing and examines different machine learning techniques that aid in the construction of trusted fog networks. This paper proposed a model that will self-train the fog node to detect malicious node in fog network.