In　recent　years,　the　Chinas　economy　has　developed　rapidly.　The　peoples　living　standard　has　been　improved.　The　number　of　cars　has　been　increasing,　and　the　pollutant　NO　has　been　produced　continuously,　which　leads　to　the　formation　of　NO2.　These　harmful　particles　have　an　impact　on　human　health.　Thus,　the　effective　and　accurate　NO2　concentration　prediction　model　plays　an　effective　role　in　peoples　health　and　prevention.　For　this　problem,　this　paper　presents　a　prediction　model　based　on　the　long　short-term　memory　(LSTM)　method　to　predict　NO2　concentration.　Firstly,　the　PM10,　SO2,　NO2,　CO,　O3,　temperature　in　a　campus　monitoring　point　in　Beijing　is　collected　as　the　research　object　in　this　paper.　Then,　the　LSTM　prediction　model　and　BP　(back　propagation)　neural　network　prediction　model　are　established　respectively.　Finally,　the　accuracy　of　the　two　prediction　models　for　the　prediction　of　NO2　concentration　is　compared.　The　results　show　that　the　prediction　model　based　on　LSTM　method　is　superior　to　BP　neural　network　model,　and　the　prediction　accuracy　is　more　accurate.　©　2021,　The　Editor(s)　(if　applicable)　and　The　Author(s),　under　exclusive　license　to　Springer　Nature　Singapore　Pte　Ltd.