Abstract: The development, testing, and evaluation of intelligent algorithms require a diverse array of testing scenarios to verify their key performance. The key difficulties in testing and evaluating intelligent algorithms are focused on, notably the scarcity and the construction complexity associated with interactive digital test scenarios. The intelligence generative capability demands of digital testing scenarios are analyzed from such four aspects as diversity, rationality, interactivity and automation capability. A technical framework for digital scenario generation based on large models is proposed, key methodologies such as generative algorithm-driven scenario layout construction, large model-enabled scenario generation intention reasoning, and layout constraint-based three-dimensional simulation scenario generation, are utilized to implement the intelligent generation of digital scenarios that meet the generative requirements. The innovative ideas and methods for the advancement of various types of command and control intelligent algorithms are provided.