This study outlines a cloud-based optimization framework for the scheduling of smart energy distribution, supported by real-time energy data from smart meters, IoT sensors, and SCADA systems. The framework uses on-board preprocessing techniques to eliminate noise, synchronize, and extract features from diverse energy datasets. The actual optimization models can be run on scalable cloud-based platforms. The authors use Mixed Integer Linear Programming (MILP) for day-ahead and intra-day scheduling optimization and Reinforcement Learning (RL) for real-time adaptive control. The intelligent scheduler uses IoT messaging protocols according to NIST standards for secure updates to grid conditions, to dynamically monitor grid states. The scheduler will push updated schedules to grid devices to align energy flow with customer demand-dispatch control. The relevant utility dashboard monitors energy utilization efficiency, cost-reduction, grid stability, scheduling latency and CO2 emission reductions. Comparison with contemporary control techniques, like time-based, rule-based, heuristic, and edge-based scheduling methods proved that day ahead and hour ahead smart scheduling in a cloud supported optimization solution is efficient and effective compared to traditional scheduling methods. This research has confirmed that cloud optimization of energy efficiency and grid resiliency can be achieved while reducing operational costs and carbon footprint emissions, representing a viable, multi-industry solution to the future smart grid complex system.