Abstract: The maintenance of integrity in bolted connections remains a challenging issue, especially when subjected to external disturbances such as vibrations, which may cause extensive or localized slip at the interface surfaces. This slip phenomenon exacerbates the relative motion between interfaces, leading to a decrease in preload levels, i.e., loosening of bolted assemblies. Over the past decade, detection methods such as vibration, guided waves and electromechanical impedance techniques have been gradually applied for detecting loosening in bolted connections. With the significant advancement in computational capabilities, machine learning algorithms including neural networks and support vector machines have been developed to further enhance the accuracy of bolt loosening detection methods. The integration of these methods offers a new pathway for real-time health monitoring of bolted connections. This paper reviews the application of methods based on the acoustoelastic effect, vibration, guided waves, electromechanical impedance, and the application of signal analysis methods based on machine learning algorithms in the field of bolt connection looseness detection and monitoring, aiming to showcase the research progress in this field in recent years.