Extended reality (XR) environments are successfully used in various domains, such as medicine, education, training, and industry. The essential element of every XR environment, apart from the XR/3D content, is the domain knowledge encoded in the environment, which is expressed by the behavior of the content, including its autonomous actions and interactions with the XR users. Such knowledge covers the processes occurring in the environment, characteristics of its objects, and skills of the users. However, the available approaches to XR creation and usage focus on the sensory perception of XR, including its visual, haptic, and aural elements, and knowledge acquisition using human senses, but they do not enable knowledge acquisition based on automated reasoning and queries. It strongly limits the possibilities of knowledge discovery from XR, in particular on a massive scale in web based environments and repositories. In this paper, we propose a knowledge-based model to 3D content behavior representation. The model permits generic representation of temporal properties of XR environments with domain terminology, which can be subject to semantic exploration with queries and reasoning. Our solution may contribute to broader and more efficient dissemination of knowledge using XR.