Abstract: The rapid growth of distributed energy resources (DER) in energy internet of things (EIoT) poses a challenge to traditional scheduling modes, which mainly cater to large-scale loads and generators. In response, our work proposes a novel framework that leverages swarm intelligence arising from the aggregation behavior of diverse DERs in the virtual space. Our framework integrates virtual twins, data science, systems theory, and 4th-Paradigm (data-intensive scientific discovery paradigm), to facilitate a cutting-edge energy scheduling approach. Concretely, we use system theory and 4th-Paradigm as the guiding ideology; we set big data analysis, machine learning, and human–machine hybrid intelligence as the core; we take digital twin, virtual simulation, high-dimensional statistics, spatial-temporal data analysis, human-in-the-loop, and knowledge embedding as the technical means. Our goal is to achieve data empowerment and intelligence improvement through seamless data connectivity, virtual-real interaction, ultimately leading to the development of a new theory on complex system scheduling.