With the advent of the Internet of Things healthcare systems, get revolutionized to offer smart healthcare services, such as remote patient monitoring, telehealth, and electronic health record systems. Electronic healthcare systems enable healthcare providers to store, manage, and access patient records electronically. However, most healthcare record systems are centralized and use conventional approaches when exchanging their data among hospitals. Due to this the security and privacy of patient’s data remains a critical challenge. To respond to this challenge, this paper introduces a pioneering framework that leverages the combined strengths of federated learning and onion routing mechanisms to fortify the security and privacy of electronic healthcare data exchange. In this context, the hospital’s local computers serve as autonomous local clients, and the hospital’s firewall acts as the role of a centralized global server overseeing the federated learning process. It offers remarkable accuracy in classifying malicious and non-malicious electronic healthcare record requests. Further, only non-malicious electronic healthcare record requests are forwarded to the onion routing mechanism, which offers anonymity and robust encryption to the healthcare record data exchange process. Moreover, for enhanced security within the onion routing network, a cryptographic verifying token is employed to validate the integrity of onion routers. These tokens are securely stored within blockchain nodes, safeguarding against tampering attempts by potential adversaries. Also, we used a 5G network interface to enhance the latency of the proposed framework. The proposed framework is evaluated with different evaluation metrics, such as accuracy, loss, de-anonymization rate, and packet error rate.