<p>This research develops an innovative autonomous battery health management framework for electric vehicles (EVs) use Edge Computing, Convolutional Neural Networks (CNNs) and Deep Reinforcement Learning (DRL), for fast, robust fault detection, adaptive charging optimization, and real-time decision making. However, traditional SoC models and simple machine learning algorithms have been sufficiently exploited in previous research for battery health monitoring making fault detection, and estimating the lifespan. But the accuracy of health predictions, the speed of fault detection and the flexibility of charging strategies that can adapt to the real-world battery condition were often marginalized by these methods. Overcoming these limitations, CNNs are used in this study to process time-series sensor data for battery degradation detection with 99.8% accuracy better than the traditional models. DRL learns in real time from the complex data of real time battery and dynamically optimizes charging cycle cycles by 55% improving energy efficiency of the energy of 35%. The system exploits Edge Computing to achieve low latency and local processing of battery health data to reduce enough state-of-charge (SoC) deviation to 6 versus 15 for baseline systems. By allowing fault detection time reduction by 40%, system reliability is increased by 50% and ensures better operational safety and efficiency. To integrate seamlessly into EVs, we develop a framework with Python and TensorFlow based, which delivers real time, autonomous solution with minimized communication overhead and optimize battery performance to support the development of more sustainable, advanced, efficient and intelligent electric vehicles.The framework monitors both electrical parameters (Current, Charge state, voltage, internal resistance) and non-electrical/electrochemical parameters (temperature, thermal gradients, degradation indices), ensuring a comprehensive view of battery health.</p>