Conventional methods for optimizing sleep, such as sleep diaries and questionnaires, frequently depend on subjective data that is vulnerable to variation and personal preference. These techniques fall short of offering precise and useful insights for individualized sleep enhancements. In contrast, this study collected wearable data from 50 participants over a 30-day period, and the analysis was driven by AI to uncover insights and provide personalized recommendations. The method examined physical activity, heart rate variability, and sleep stages using sophisticated machine learning models, particularly Long Short-Term Memory (LSTM) networks. The association between environmental conditions and individual sleep behavior over time revealed a 12% reduction in nighttime disturbances and a 20% increase in deep sleep duration. The suggested approach outperforms conventional methods by providing unbiased, data-driven therapies customized to the unique sleep habits of each person. The AI-driven system showed a considerable improvement in sleep quality through focused interventions, making it a superior alternative for optimizing sleep compared to conventional approaches that lack accuracy and customization. © 2025 Elsevier B.V., All rights reserved.