Sleep disorders have become a global public health problem, and traditional diagnosis and treatment are limited by low accessibility and high subjectivity. In recent years, artificial intelligence (AI) technology has brought innovative opportunities for sleep health monitoring, diagnosis and intervention. In this paper, we systematically review the research progress of AI in the field of sleep: home monitoring based on wearable devices and contactless sensing achieves continuous and senseless data collection; deep learning models (such as convolutional neural network, recurrent neural network, and their hybrid architectures) significantly improve the automation and accuracy of sleep staging; and in clinical applications, AI assists in the diagnosis of obstructive sleep apnea, promotes the development of OSA diagnosis and treatment technologies as well as related research, and supports to form personalised intervention strategies. Despite its bright prospects, the application of AI in sleep health still faces numerous challenges, such as data privacy, algorithmic bias, insufficient model interpretability and lack of clinical validation. It should be made clear that AI is positioned to enhance, rather than replace clinical professional judgment. In the future, multimodal fusion, generative AI and big language models for health will further contribute to the intelligence and personalisation of sleep management. This study aims to provide reference for researchers and clinicians to promote the deep integration of AI and sleep medicine.