Simultaneous　localization　and　mapping,　SLAM　can　product　a　Map　for　autonomous　robots　and　self-driving　vehicle　in　navigation.　In　the　actual　environment,　the　scene　changes　frequently,　which　makes　the　old　map　no　long　reliable.　Therefore,　it　is　necessary　to　update　such　a　map　by　an　efficient　and　safely　way.　In　this　paper,　we　review　the　existing　map　updating,　long-term　localization　methods　and　discuss　about　the　challenges　in　this　situation.　We　present　a　Map　updating　method　in　mathematical　way　which　can　update　accurately.　Our　proposed　method　are　tested　in　five　indoor　dataset　and　demonstrated　feasibility.　©　2021　IEEE.