Abstract: Complexity is an important feature and trend of the energy Internet of things (EIoT), which may disable the simulation, analysis, and deduction methods based on traditional mechanistic models (also regarded as white box) in EIoT—a complex system essentially. The integration of data science, simulation technology and virtual twin is being advocated as a powerful tool to deal with complex systems. Aiming at the complexity, this paper explores the virtual twin technology and its promising applications. Focusing on the basic properties and technical roadmap, a cognition framework of emergence phenomena in complex systems based on virtual simulation deduction technology is proposed. This work provides a reference for carrying out virtual simulation task in EIoT on the one hand, and is beneficial to the advancement of virtual simulation technology on the other hand.