Internet of Vehicles (IoV) is transforming transportation today using artificial intelligence (AI) and blockchain to improve road safety. However, driver behavior, collision risks, and data forgery remain essential issues. In this paper, a Blockchain-Based Trust and Incentive Model (BTIM) is presented to endow vehicles with trust scores based on their real-world driving behaviors. A machine learning-based algorithm estimates collision risks based on speed, braking habit, and inter-vehicle gap. The blockchain ledger provides immutable trust scores, and smart contracts automate rewards for safe driving and punishment for dangerous driving. Experimental results show that BTIM realizes 99.4% collision prediction accuracy while providing secure, decentralized trust management with 50% less processing latency. These results identify BTIM as an efficient and scalable solution for trust-based safety in IoV ecosystems.