Abstract: In response to the problem of safety monitoring of ancient building structures under tunnel construction disturbance, this study focuses on the Baita Mountain building complex in Lanzhou and constructs a BIM based monitoring system for ancient building structures. Through 3D laser scanning and point cloud reverse modeling technology, the millimeter level ancient building information model covering tower base, column grid and bracket sets (Dougong) of wooden architecture nodes is established to provide a high fidelity digital base for structural analysis. We independently develop a BIM family library that includes information such as size, stress, price, material, installation sequence, etc., and combine CAD axis positioning technology to achieve accurate positioning of BIM models. Based on the structural characteristics of ancient buildings, a monitoring scheme of "minimally invasive implantation + multi-source fusion" is proposed to construct a three-level monitoring system. Build a 3D visualization intelligent monitoring platform based on machine learning algorithms to achieve dynamic display, safety warning, and multi-dimensional comparative analysis of monitoring data. The monitoring system developed in this study has broken through the problem of data silos in traditional monitoring models and formed a scalable intelligent monitoring technology paradigm for ancient architectural complexes, providing important technical support for the digital protection of historical buildings and the coordinated implementation of major projects.