Urban drainage systems face increasing challenges due to rapid urbanization, climate variability, and aging infrastructure. Traditional drainage modelling approaches, often reliant on static data and computationally intensive hydraulic simulations, struggle to meet the real-time demands of modern flood management. While recent studies have explored data-driven models, graph neural networks, and surrogate simulations, many remain limited by insufficient sensor integration, lack of adaptive control, and restricted real-world applicability. This paper proposes an integrated real-time urban drainage modelling framework that leverages continuous sensor data assimilation, intelligent flood mitigation algorithms, and adaptive infrastructure control. By combining live data streams from IoT-based sensors with advanced predictive modelling and dynamic control mechanisms, the system enables timely, accurate decision-making for flood prevention and optimized drainage operations. The proposed approach addresses key gaps in existing research by enhancing system scalability, computational efficiency, and multi-hazard resilience while supporting the development of digital twin ecosystems for urban water infrastructure. Experimental validation demonstrates the effectiveness of the framework in improving operational responsiveness and ensuring sustainable, adaptive flood management in smart cities.