Based on AI and machine learning, this book systematically presents the theories and methods for complex electro-mechanical system fault diagnosis and health state assessment in modern industry. The book emphasizes feature extraction, incipient fault prediction, fault classification, and degradation assessment, which is based on the supervised, semi-supervised, manifold, and deep learning, machinery degradation tracking and prognostics based on the discriminative features extracted by phase space reconstruction, and complex electro-mechanical system reliability assessment and health maintenance based on running state info. These theories and methods are integrated with practical industrial applications, which can help the readers get into the field more smoothly and provide an important reference for their study, research, and engineering practice.