To bridge the gap between “Analyze-Decide-Act” chain, we extend the software-defined decision execution component to extend traditional decision support systems, define the Software-Defined Decision-Making Operating System (SD-DMOS), and propose a generic concept of Decision-Making Operation Space as well as a prototype architecture of SD-DMOS based on dynamic ontology. To effectively shorten the time consumption of the “Observe-Orient-Decide-Act” loop based on heterogeneous multi-modal big data analysis, and enhance the efficiency of collaborative analysis, decision-making, and execution under the premise of security and controllability, we explore software-defined dynamic ontology construction, automatic deployment and management of software-defined actuators, and software-defined data space perimeter technology. These efforts lay a solid foundation for further research and system design implementation.