Prognostics and health management (PHM) of power transformers is of great significance for realizing the transition from traditional regular maintenance to condition-based maintenance and predictive maintenance to ensure the healthy operation. For a long time, the transformer PHM technology has stayed in the theoretical research stage, and there is a lack of an effective technical system and platform to integrate and improve the performance of the research results of each stage. The digital twin (DT) technology strengthens the monitoring and integration of the operating parameters of the transformer’s multi-physical components. Through multi-physical and multi-scale modeling in the virtual space, it realizes comprehensive fault analysis for transformers and has become an important direction for the evolution of transformer PHM. In view of this, the DT technology for power transformer PHM is studied, the connotation is explained, and the DT technology framework, key technologies, challenges and future development trends are summarized. The relationships among DT and its supported data, models, calculations, etc. are analyzed. It aims to provide references for researchers of digital twin technology in the field of transformer operation and maintenance.