The intelligent construction of airports faces critical challenges, including high BIM data heterogeneity, insufficient accuracy in dynamic construction prediction, and low efficiency in multi-source sensor coordination, which hinder engineering intelligence. This study proposes a data governance-driven AI-BIM integration framework. By developing a metadata standardization engine based on DAMA 2.0, data preprocessing efficiency is improved by 3.8 times. An enhanced ResNet3D-50 architecture incorporating a channel attention mechanism (CBAM module) significantly enhances multi-scale feature extraction for BIM point clouds, achieving a recognition accuracy of 93.4±0.7% while reducing model parameters by 62%. Real-time data acquisition via IoT sensors, combined with dynamic BIM model updates, enables efficient and safe airfield construction without disrupting flight operations. The results demonstrate that this system substantially improves construction efficiency and quality This research provides theoretical foundations and practical guidance for intelligent airport construction, offering significant application value and broad promotion potential.