Abstract: Mechanical energy harvesting is a process of converting the scattered, disordered, low-quality, and high entropy mechanical energy from the environment into electrical energy. It enables self-powered IoT by powering low-power electronic devices such as widely distributed sensors. Mechanical energy harvesting exhibits the merits of flexible, convenient, sustainable, zero carbon, and environmental-friendly. It has become an international frontier research hotspot and can be widely applied in natural environment monitoring, infrastructure condition monitoring, equipment condition monitoring, etc. However, key issues are restricting the practical application of mechanical energy harvesting, such as low output power, narrow working frequency band, poor low-frequency effect, poor environmental adaptability, and low reliability. The dynamic regulation method for mechanical energy harvesting has the potential to enhance the performance of such systems by tailoring them to specific environmental stimuli and improving their output electrical performance. This paper constructs a dynamic regulation methodology system including excitation regulation, nonlinear systems, multi-degree of freedom systems, adaptive control and strategy regulation. The latest advancements in dynamic regulation methods has been discussed, including the characteristics and typical designs of each type of dynamic regulation method. Finally, the main challenges faced by dynamic regulation methods were summarized and the development trends were discussed. This paper provides a new perspective for adaptive dynamic regulation of mechanical energy harvesting systems in complex environments, which is beneficial for promoting the development of mechanical energy harvesting theory and technology.