Machine learning (ML) and the Internet of Things (IoT) are revolutionizing healthcare through real-time monitoring and predictive analytics. This study presents HealthGuard, an AI-driven remote health monitoring system designed to enhance patient care and reduce emergency response times. It integrates wearable sensors, an ESP32 microcontroller, and a cloud-based infrastructure to collect vital signs like heart rate, blood pressure, SPO \(_{2}\) , and body temperature. Data is securely transmitted to the cloud, where a random forest algorithm detects anomalies. Upon identifying deviations, HealthGuard sends real-time alerts via mobile and web applications, enabling prompt medical intervention. Its modular design supports additional sensors and advanced ML models, making it adaptable to various healthcare needs. Experimental results indicate high accuracy in anomaly detection, particularly benefiting elderly patients with chronic conditions. By facilitating continuous monitoring and early detection, HealthGuard improves patient outcomes and reduces the burden on healthcare systems. This paper assesses its accuracy, performance, and usability.