This research focuses on the use of ML models in real-time patient monitoring systems for IoT-enabled healthcare environments. Data from one month of sensor measurements of different devices, such as ECG, SpO2, body temperature, accelerometer, and blood pressure sensors, are collected, and 3500 records were produced. The data was split into training (70%) and testing (30%) sets, and machine learning (ML) models like artificial neural network (ANN), support vector machine (SVM), decision tree (DT), and recurrent neural network (RNN) were used to predict the health conditions of patients. The performance of each model was measured by applying the appropriate metrics, precision, recall, and F1-score. The ANN model had the highest precision at 97.66%, followed by SVM at 93.45%, then DT at 87.67% while RNN results were last at 85.54%. Apart from the accuracy achieved, inference time and memory usage were considered. ANN possessed a decent balance between attaining accuracy and using resources. The models are installed in real-time healthcare monitoring systems that collect sensor data, process, and make a prediction for instant alerts concerning abnormal conditions. It gives real-time and accurate information to health providers, making them adopt proactive health management. The results of this study indicate that the ANN is best suited to high-precision applications while offering DT as an alternative that performs faster and uses fewer resources. This study demonstrates the potential of IoT and ML integration for enhancing healthcare delivery through continuous real-time monitoring and early intervention.