Vehicular Ad-hoc Network (VANET) is a pivotal factor in an Intelligent Transportation System (ITS) that improves traffic management, road safety, and infotainment features. Routing in VANET is impeded by frequent link disconnection due to roadblocks, high mobility, and rapid topological changes. An Unmanned Aerial Vehicle (UAV) with three-dimensional movement capacity may significantly enhance a VANET’s routing experience by boosting the likelihood of a line-of-sight, improving connection, and implementing an effective store-carry-forward mechanism. This also enables efficient data transmission by minimizing free-space path loss, a key contributor to signal degradation. The proposed approach addresses connectivity and congestion issues in vehicular networks by utilizing UAVs as relay nodes and implementing a congestion-aware routing protocol based on Q-learning. This method leverages vehicle density patterns, network state monitoring, and hierarchical packet routing to ensure continuous connectivity, reduce communication overhead, and enhance packet delivery reliability. The results show that the proposed approach significantly enhances packet survivability and packet delivery ratio (PDR).