[Purposes] In order to obtain more intuitive and accurate original data of tunnel surrounding rock, a three-dimensional reconstruction method of tunnel surrounding rock based on structure from motion (SfM) was proposed. [Methods]First, the scale-invariant feature transform (SIFT) algorithm was used to extract feature points from the tunnel surrounding rock images. Second, Euclidean distance feature matching on these feature points was performed with the SfM principle to achieve camera parameter calibration and generate sparse point clouds of the tunnel surrounding rock. Third, the multi-view stereo (MVS) algorithm was applied to densify the sparse point cloud so as to obtain a dense point cloud of the high-density surrounding rock. Finally, based on the Delaunay triangulation algorithm, surface reconstruction and texture mapping on the dense point cloud were completed to generate a three-dimensional model of the tunnel surrounding rock. [Findings]More than two hundred real-shot images of the tunnel surrounding rock were used for experimental verification. The results show that this method can effectively reconstruct a complete and realistic three-dimensional model of the tunnel surrounding rock. [Conclusions] This method demonstrates high engineering applicability and popularization potential and provides a new technical idea and reference for data collection of tunnel surrounding rock and refinement analysis of geological information.