This research explores the functional analysis and prediction of future states in 3D indoor scene environments using textual descriptions. The focus is on understanding how indoor 3D point cloud environments can be dynamically altered based on specific text prompts. This involves predicting changes in object placement, room configuration, and scene functionality, which is crucial for applications in robotics, virtual reality, and smart home systems. By leveraging text-based prompts to predict future states of the environment, the study aims to reduce the dependency on real-time scene reconstruction from RGB-D frames, enhancing efficiency in tasks like automated furniture arrangement, object manipulation, and environment adaptation in response to user commands. The broader objective is to create intelligent systems that can interpret and act on human language to modify and interact with 3D environments in a functionally meaningful way.