The　energy　consumption　of　large　airport　terminal　buildings　is　twice　or　even　higher　than　that　of　general　large　public　buildings.　The　energy　consumption　prediction　can　obtain　the　energy　operation　law　of　the　building　in　advance.　Furthermore,　the　unreasonable　behavior　of　using　energy　of　buildings　can　be　interfered　in　advance　to　save　energy.　This　study　takes　a　large　airport　terminal　in　Beijing　(China)　as　the　object.　According　to　the　power　consumption　characteristics　of　its　main　electrical　equipment　system,　combined　with　Pearson　correlation　analysis,　multiple　regression　analysis　and　neural　network　methods,　the　influence　of　passenger　number　and　outdoor　meteorological　parameters　on　the　power　consumption　and　their　correlation　characteristics　are　studied.　The　results　show　that:　(1)　The　number　of　passengers　and　outdoor　temperature　are　the　key　factors　affecting　the　power　consumption　characteristics　of　airport　terminals;　(2)　Compared　with　the　regression　analysis　model　and　the　BP-GA　neural　network　model,　the　prediction　accuracy　of　LSTM　neural　network　model　is　higher.　Its　RMSE　is　1917.0072,　MAPE　is　0.49　%,　R2　is　0.99;　(3)　LSTM　neural　network　model　is　more　suitable　for　the　learning　and　training　of　the　power　consumption　law　of　airport　terminals.　©　2024　18th　Conference　of　the　International　Society　of　Indoor　Air　Quality　and　Climate,　INDOOR　AIR　2024　-　Conference　Program　and　Proceedings.　All　rights　reserved.