The high penetration of power electronic converters in modern power systems has led to increasingly complex electromagnetic transient (EMT) dynamics and stronger dynamic coupling

posing significant challenges to traditional EMT simulation in terms of computational efficiency and scalability. Network partitioning and decoupling techniques are crucial for enhancing simulation performance of large-scale power electronic systems. This paper systematically reviews recent advances in EMT partitioning and decoupling methods from both domestic and international research. The mainstream approaches are categorized into three types based on their modeling approach of interfaces and earliest proposed time: delay-based models

interface equivalent models

and delay-free models. Comparative analyses are conducted using typical case studies to evaluate their performance in accuracy

computational efficiency

and interface stability. Based on this analysis

the applicability and limitations of each category are examined. Future research directions are highlighted from three perspectives: modeling improvements (e.g.

multi-method fusion and enhanced stability); integration with emerging technologies (e.g.

digital twin

artificial intelligence

and quantum computing); and engineering deployment and standardization. This review provides theoretical insights and technical guidance for efficient simulation of large-scale power electronic systems

supporting the ongoing development of high-accuracy

high-efficiency

and practical solutions.