This research focuses on developing and implementing an active-adaptive construction project management system based on Kohonen Self-Organizing Maps (SOM) technology. The high variability of architectural and structural solutions, complex design dynamics, and multifactorial engineering calculations in modern construction necessitate creating flexible automated management systems capable of self-regulation. The research methodology integrates cluster analysis of design characteristics, multidimensional topological mapping of structural elements, and neural network analysis using SOM algorithms. The empirical base encompasses data from 38 construction projects of various scales during 2019-2023, with a total area exceeding 4.3 million square meters. Results demonstrate a 36.4% reduction in design documentation development time, 21.7% decrease in structural material consumption, and 17.3% improvement in building energy efficiency. A strong correlation (r=0.83) was established between the degree of structural solution optimization and economic efficiency of construction projects. The developed system provides dynamic visualization of multi-parameter design solution structures, enabling real-time identification of critical contradictions and preventive correction of potentially problematic structural nodes. The research significance is confirmed by multifactorial economic implementation efficiency (ROI=2.7) and substantial reduction in construction timeframes (average 14.6%).