Ransomware attacks have increasingly shifted toward more sophisticated tactics, not only encrypting critical files but also exfiltrating sensitive data, which is then used as leverage in extortion attempts. Addressing this dual threat, a novel framework integrating differential privacy provides an enhanced layer of protection by ensuring that exfiltrated data remains unusable to attackers through the introduction of statistical noise. This approach uniquely combines differential privacy with traditional security techniques, allowing for a dynamic and adaptable defense mechanism that ensures both data utility and robust privacy guarantees during ransomware attacks. The framework effectively mitigates data theft by applying controlled noise to sensitive datasets, which significantly reduces the probability of successful re-identification, even when auxiliary information is available. Through experimental evaluation, the framework has demonstrated superior performance in balancing privacy, utility, and system efficiency when compared to existing ransomware defense mechanisms. Furthermore, the modular nature of the system allows for seamless integration into existing cybersecurity infrastructures, ensuring that organizations can implement the solution without major architectural changes. Overall, the proposed framework offers a promising advancement in protecting sensitive data from both encryption and exfiltration threats, providing a comprehensive and adaptive approach to modern ransomware challenges.