In the rapidly evolving landscape of electric vehicles (EVs), integrating the Internet of Vehicles (IoV) has opened up new dimensions in automotive technology. While enhancing vehicle performance, safety, and security, this advancement also exposes EVs to heightened cyber threats. To address this critical concern, we introduce GUARDEV: Guided Unified Adaptive Response for Defending Electric Vehicles, a novel Machine Learning (ML)-enhanced conflict resolution-based ensemble Intrusion Detection System (IDS) architecture. GUARDEV leverages the power of adaptive ML techniques to provide a unified and guided response in safeguarding EVs against sophisticated cyber-attacks. Our framework evaluates the effectiveness of three prominent ML models - XGBoost, Random Forest, and Extra Trees - in detecting specific categories of cyber threats. By harnessing the prediction confidence scores from these models, GUARDEV offers a nuanced and intelligent approach to recognizing and thwarting diverse cyber-attacks. Our research, grounded in rigorous experimentation using public IoV security datasets, demonstrates GUARDEV’s exceptional accuracy of 99.47% in intrusion detection. This remarkable achievement not only underscores the potential of ML in enhancing vehicular cybersecurity but also establishes GUARDEV as a powerful tool for creating safer, more secure EVs. GUARDEV paves the way for robust protection of internet-connected electric vehicles in today’s digital age by providing a guided, unified, and adaptive response to cyber threats.