<p>Health prediction is a very popular application of wearable devices in the current scenario. There is always a need for advanced technology in this area. Wearable technologies such as the Apple Watch and Fitbit have transformed from basic fitness trackers into advanced platforms for continuous health monitoring and prediction. These devices generate granular, multi-dimensional data streams encompassing physiological metrics like heart rate variability, physical activity, caloric expenditure, sleep patterns, and stress indicators. The need for highly precise automated models motivated researchers to develop the artificial intelligence framework for health predictions. When coupled with state-of-the-art artificial intelligence (AI) and machine learning (ML) methodologies, this data enables the construction of predictive models capable of early health risk detection, personalized intervention, and improved health management. This study offers a detailed examination of modern analytical techniques applied to wearable sensor data, focusing particularly on datasets from Apple Watch and Fitbit. A comparative analysis is conducted across various domains, including health prediction, activity recognition, sleep analysis, and stress detection. Furthermore, a bibliometric study highlights key trends, research evolution, and emerging focus areas in this field. The paper concludes by proposing future research directions, emphasizing the integration of advanced AI frameworks and ethical considerations to optimize personalized health recommendations and ensure responsible deployment.</p>