The complexity of traditional data mining techniques prevents them from being used in VANETs. The introduction of data stream mining has opened a window for extending data mining to be used in VANETs. Data stream mining can be used to construct or extract patterns from messages broadcast frequently from vehicles traveling on a road. Vehicles exchange different types of messages either periodically or ad hoc for different types of applications. The data flowing in the network of vehicles can be used to extract valuable knowledge to support the different applications in VANETs. Knowledge gained from the data can be used to form groups of vehicles. Vehicles can be grouped together based on their direction of traveling, speed, the types of applications they run. Vehicles that have common features, interests, and needs facilitate the establishment of trust between them as these shared features make up the foundation for trust. In this paper we provide a discussion of establishing trusted group using trust rules. In addition, we propose a system to generate/process trust rules and to establish a set of trusted groups.