Vehicular networks have become a fundamental component of next-generation intelligent transportation systems, with the Internet of Vehicles (IoV) enabling communication among vehicles and infrastructure. While this connectivity offers significant benefits, the large scale of participating vehicles, the high frequency of message exchanges, and inherent vehicle mobility introduce substantial challenges to ensuring strong security, privacy preservation, and trust management during data sharing. This thesis proposes novel Reinforcement Learning (RL)-based approaches supported by immutable blockchain infrastructure to enhance security, privacy, and trust (SPT) provisioning in vehicular networks. To structure the analysis of existing approaches, two distinct taxonomies are proposed in this thesis: a focus-centric taxonomy and a taxonomy based on two dimensions: the type of SPT protection and the type of the RL methodology employed. To enhance trust management, two complementary RL-based approaches are proposed: a link-centric Q-learning-based Trustworthy Routing Technique (QL-TRT) that selects trusted routes while strengthening availability, and a vehicle-centric Dynamic Selection of Trusted Seller using Deep Deterministic Policy Gradient algorithm (DSTS-DDPG) that dynamically optimizes credibility score thresholds for trusted data provider selection in vehicular edge computing (VEC). To enhance privacy preservation, a Blockchain-enabled Pseudonym Management for VEC (BPMV) framework is proposed, integrating a permissioned blockchain with a hierarchical pseudonym structure, together with a Cost-Aware Pseudonym Management algorithm using Multi-Agent Proximal Policy Optimization (CAPM-MAPPO) that balances vehicle-side and VEC-side costs against privacy entropy. To address privacy and trust jointly, a Blockchain-based Reputation Management for Privacy Preservation (BRMPP) framework is proposed, together with a Trust-Aware Pseudonym management algorithm based on Proximal Policy Optimization (TAP-PPO) that enforces reputation constraints on the policy space through a trust-aware action-masking mechanism, ensuring reputation continuity across pseudonym rotations in different Certificate Authority regions. Extensive simulations were conducted across diverse scenarios to evaluate the proposed solutions. Average reward was assessed for all RL-based approaches, while additional metrics including privacy entropy, packet delivery ratio, energy consumption, and cost were applied selectively based on each contribution’s specific objectives. The results collectively validate the effectiveness of the proposed algorithms in real-time vehicular environments and highlight their potential for large-scale deployment.