The rapid adoption of 5G-enabled Internet of Things (IoT) devices in urban transportation networks has enabled smarter, more efficient mobility systems. However, this interconnected infrastructure introduces critical cybersecurity vulnerabilities, especially for low-power IoT devices deployed at the edge—such as traffic sensors, roadside units, and connected vehicles. In this paper, we propose a lightweight cybersecurity framework tailored for 5G-enabled urban mobility environments. The framework integrates secure communication protocols, device authentication, and anomaly detection using low-resource machine learning techniques. Our approach balances security with computational efficiency, ensuring suitability for constrained devices without sacrificing resilience to common cyber threats such as spoofing, denial-of-service, and data tampering. We present a modular architecture adaptable to heterogeneous IoT deployments and evaluate its performance through simulation and scenario-based testing. Results show improved detection accuracy and latency performance compared to baseline methods, validating the framework’s feasibility for real-world smart city applications. This work contributes toward building secure-by-design infrastructures for future cities, where mobility and communication are increasingly intertwined.