Great difficulties are encountered in the modern smart grid system when aims of reducing losses, stabilizing voltage and properly incorporating renewable energy resources need to be achieved for a dynamic operation of the system. Current optimization methods primarily target single-objective control strategies and do not include adaptive preventive learning features to support real-time control of smart grid. To overcome these drawbacks, this research is proposing a Preventive Synaptic Learning Smart Grid Optimization (PSLSGO) framework to optimize the utilization of smart energy management and power distribution. The proposed adaptive framework brings together adaptive synaptic learning, preventive fault prediction, intelligent power routing, reactive power compensation, adaptive load balancing and renewable energy coordination in a single optimization framework. The real-time electrical parameters such as voltage, current, frequency, load demand, transformer temperature, and renewable energy generation were gathered using IoT-enabled monitoring systems and analyzed with the help of the IEEE 33-bus smart distribution network. The experimental results showed that the proposed PSLSGO framework was able to achieve 98.4% energy efficiency, 98.1% reliability, 48.9% reduction in transmission losses, 97.2% voltage stability and 92.5% renewable energy utilization. The obtained results were statistically validated using ANOVA and confidence interval analysis to ensure the significance and reliability of the results. The adaptive synaptic learning mechanism has been found to enhance the prediction accuracy and capability of preventive control, in the presence of varying operating conditions. In conclusion, the proposed PSLSGO framework offers a smart, scalable, and energy-efficient approach for future smart grid optimization and sustainable power distribution system.