Abstract: To overcome the shortcomings of traditional subpixel mapping methods based on image fusion, which cannot fully extract the spatial information of panchromatic images, in this study, we propose a subpixel mapping method based on superpixel segmentation and panchromatic image fusion. First, the spatial information of the panchromatic image is extracted using superpixel segmentation. Then, guided by the superpixel image, the subpixel abundances obtained by unmixing are mean-fused within the superpixel regions to obtain the superpixel abundances. Finally, the homogeneity of the superpixel region is assessed according to the component complexity of the superpixel region, and the winner-take-all strategy is used for the homogeneous superpixel region. The pixel swapping algorithm at the superpixel scale is used for the nonhomogeneous superpixel region, and an improved subpixel mapping result is obtained. The proposed method is validated using two commonly used remote sensing image datasets, and the results demonstrate that compared to most traditional methods, the proposed method can further enhance the subpixel mapping accuracy while maintaining smoothness and better preserving details.