There are high demands for the early and reliable detection of metal components used in safety-critical structures. Nondestructive testing (NDT) is a pivotal technique used across industries to assess a material’s integrity without causing damage and has been used in early crack detection of metals, mainly based on changes in the crystal structure and magnetic properties of metals. This review provides an overview of internal and external detection technology based on nondestructive testing methods such as ultrasonic, electromagnetic, ray, magnetic particle, etc. Especially, the integration of advanced methodologies such as machine learning and artificial intelligence deserves a place in NDT methods. Furthermore, the multifactorial detection method is promoted to enhance the sensitivity and detection range due to advantage integration but still has emerging challenges for safer equipment and applications. The review aims to compare these methods and outline the future challenges of NDT technologies for metal crack detection.