The invention relates to the field of point cloud indexing, in particular to a method for indexing scattered point cloud space of a historic building. The method is characterized by including steps of acquiring a data construction point set; gridding the point set, setting grid index digit groups, and calculating value of each unit of the grid index digit groups according to point coordinates in grids; merging the grids into an area or dividing an area by an octree, setting area digit groups, storing ID (identification) of each area and IDs of the areas of all adjacent six sides into the corresponding digit group, constructing a K-D tree in each area, traversing grid index digit groups, and calculating IDs the areas corresponding to the grid in the area digit groups; looking for the grids and the areas where inquired points locate according to coordinates of the inquired points, indexing in the areas and the adjacent areas by means of the K-D tree and obtaining the point nearest to the inquired point. The method solves the problems of single resolution ratio and data redundancy, and guarantees inquiry efficiency of the K-D tree.