Abstract: Aiming at the problems of poor mapping quality and narrow pathways input limit in existing methods for computing real walking maps, a conformal mapping based methods is proposed for large-scale open scenes. First, some points are specified to allow the map of virtual scenes to circle around and small parts around these points are removed from the map. Then, a conformal mapping is built to make the mapping result circle around the points. The number of circles can also be specified by people. Finally, an optimization for isometric energy with the conformal mapping as initialization is adopted to further reduce the distortion throughout the map. The proposed method was applied to virtual maps with different sizes and shapes. The results of experiment show that the proposed method generates the mapping from large-scale open scenes to small physical spaces with lower isometric distortion when compared with the direct optimization method.