In　order　to　reduce　the　operation　instability　of　rail　transit　station,　optimize　the　structure　of　rail　transit　network　and　improve　the　efficiency　of　network　operation,　the　stability　of　rail　transit　operation　under　the　impact　of　large　passenger　flow　was　studied.　Taking　Beijing　rail　transit　network　as　an　example,　Beijing　rail　transit　network　model　with　rail　stations　as　nodes　and　adjacent　stations　as　edges　was　constructed　by　using　Space　L　method.　The　characteristic　parameters　and　distribution　rules　of　node　degree,　average　path　length,　clustering　coefficient,　network　diameter　and　network　efficiency　were　simulated　and　analyzed.　According　to　the　type　of　accidents,　these　stations　were　divided　into　site　operation　weakened　and　operation　disrupted,　respectively.　Base　on　passenger-flow　propagation　model　and　network　efficiency　model,　the　simulation　quantified　the　network　stability　under　the　background　of　different　event　types.　The　results　show　that　Beijing　rail　transit　network　is　a　scale-free　network.　The　accessibility　of　rail　transit　network　is　better　and　clustering　coefficient　of　it　is　small,　but　connectivity　of　it　should　be　improved.　And　when　site　operation　is　weakened　because　of　the　large　passenger　flow　in　Beijing　transit　network,　if　the　number　of　adjacent　station　is　not　more　than　7,　passenger-flow　propagation　state　in　network　can　regress,　or　negative　correlation　is　observed　between　the　number　of　adjacent　station　and　the　fading　time.　When　transfer　station　is　disrupted,　the　value　of　network　efficiency　will　decrease　significantly.　The　importance　coefficient　of　the　damaged　site　has　the　greatest　impact　on　network　efficiency　in　the　range　of　(0.75,　1],　and　the　least　in　the　range　of　(0,　0.　5].　The　research　results　provide　a　basis　for　timely　finding　the　key　points　and　weak　links　of　rail　network,　ensuring　the　safety　operation　of　urban　rail　transit　network　under　the　background　of　super　passenger　flow,　and　providing　a　decision　support　for　the　development　of　emergency　plan　of　Beijing　rail　transit　network　under　the　impact　of　large　passenger　flow.　5　tabs,　5　figs,　16　refs.　©　2018,　Editorial　Department　of　Journal　of　Chang＇an　University　(Natural　Science　Edition).　All　right　reserved.