Under the strong promotion of national policies, prefabricated buildings have ushered in a rapid development era. Compared to traditional building forms, prefabricated buildings can effectively solve problems such as labor shortage and environmental pollution in the construction industry. In this regard, this article provides corresponding optimization suggestions for the construction of prefabricated steel structure buildings based on K-Means clustering algorithm, hoping to provide reference for relevant smart building design. The research results indicate that in the process of importing horizontal components such as steel beams into the steel structure of the algorithm in this paper, the main or secondary beams that intersect with the secondary beams are merged into a complete steel beam, and the number of steel beams will be correspondingly reduced. However, in the structural analysis and calculation using the AGA (Adaptive Genetic Algorithm) and ML (Machine Learning) algorithms, there is a problem of members breaking at nodes, and the number will correspondingly increase. It can be seen that the algorithm in this article can use 3D model data to answer questions and solve doubts about the corresponding technology when managing the construction site, and the probability of problems occurring will also be greatly reduced. At the same time, a comprehensive analysis and systematic inspection will be conducted internally, and the K-Means clustering algorithm will be used to fully understand the stress situation and the connection situation of corresponding points. At the same time, detailed location problems will be solved to ensure the smooth and stable construction process.