This paper introduces a sustainable route optimization framework for autonomous vehicle (AV) fleets, integrating Internet of Things (IoT), edge computing, and digital twin (DT) technologies. The framework dynamically adjusts routes in real time, responding to traffic and environmental conditions. Utilizing multi-objective optimization algorithms and machine learning (ML), it effectively minimizes carbon emissions, fuel consumption, and travel times while enhancing overall traffic efficiency. The proposed system incorporates Transit Signal Priority (TSP) for optimized intersection management and an Environmental Impact Monitoring System for continuous evaluation of emissions and sustainability benchmarks. Designed for scalability, this model addresses critical urban mobility challenges, offering practical insights for policymakers and planners to foster greener, smarter cities.