E-Health system should provide effective, efficient service delivery to citizens anywhere at any time with the help of information and communication technologies such as the internet, local area network, mobiles, etc. The major things for a better health service system are system availability, reliability, per formability, and security. This paper proposes a proactive fault handling approach to enhance the fault tolerance level of an E-health monitoring system with a machine learning-based model to predict the upcoming fault in the system. It trains 8 different models using 8 different Machine Learning Classifiers and compares those models using 6 evaluation matrices: Accuracy, Mean Absolute Error, Precision, Recall, F-Measure and False Positive Rate. The model with better performance among all is recommended for developing Proactive Fault Handling based E-health Monitoring System. Proposed Systems with Proactive Fault Handling (PFH) and Systems without using Proactive Fault Handling are modeled using a Continuous Time Markov Chain (CTMC), and the Availability, Downtime, and Reliability of the modeled systems are computed and compared.