The project aims to merge the Internet of Things with machine learning to come up with a reliable system for heart activity monitoring and real-time arrhythmia identification. This system uses ECG sensors, temperature sensors, heart rate monitors, and a single-board computer Raspberry Pi. Using these parts, it collects real-time physiological information to be processed through a trained neural network model that gives out abnormal heart rhythms. The performance metrics indicated that the model achieved high efficiency because it obtained an AUC of 0.983 from the training data and of 0.950 in validation in terms of distinguishing normal from abnormal heartbeats. The accuracy rates are 96.3% for training data and 94.5% for validation data, underscoring the model’s reliability. Although it successfully detects irregular heartbeats, there is potential to enhance precision and reduce false positives, as indicated by precision scores of 80.8% and 73.8%. This all-encompassing solution addresses shortcomings in previous heart health monitoring systems and underscores the importance of continuous improvement. In the next phase of work, we will optimize model performance, and explore new methods even further to increase precision and robustness. Obviously, no question exists about the power of the model; but, for an improved model, the precision in identifying false positives needs to be just a little higher, as this will allow us to decrease that number and then improve on both sides a little more. Due to the real-time monitoring and several sensors, this system acts as an all in one solution for controlling of heart health which was missing from previous systems. This paper focuses on enhancing the model performance and future efforts will be spent to explore other methods that are able to further refine accuracy as well stability of the system.