The general land use planning is to arrange all kinds of land use behaviors in an orderly manner, and its own disorder will inevitably affect the scientific and authoritative nature of the general land use planning, which makes the effect of land management unsatisfactory. The reason is that there are differences in the spatial scale of general land use planning, which makes the analysis of land information not comprehensive enough. Therefore, this paper puts forward the dynamic planning of sustainable land use based on GIS and symmetry algorithm. The collected land information is preprocessed by ArcGIS, and the land remote sensing images are classified by decision tree combined with ENVI (Visual Image Environment), using local sparse coding to extract land use features, with enhancement and reconstruction of remote sensing images using symmetric algorithm. On this basis, the factors that limit the spatial expansion of land are analyzed, and the land planning is completed by substituting them into Arc Map software. Experiments show that the average accuracy rate of this method for land planning reaches 91.22%, which can effectively complete land planning.