Metro　has　already　become　one　of　the　main　transportation　methods　in　big　cities　across　the　world.　With　the　development　of　metro　systems,　lots　of　associated　safety　issues　(e.g.,　fire　accidents　and　terrorism　attacks)　have　occurred　in　the　past　decades.　Under　such　circumstances,　studies　on　the　safety　evacuation　of　the　metro　station　hall　are　very　necessary.　Due　to　high　human-density,　fluent　evacuation　is　important　in　the　metro　station　hall　in　emergencies.　However,　metro　station　hall　evacuation　is　a　complex　process,　filling　with　humans　of　different　behavior　biases.　Selection　of　evacuation　direction　usually　is　influenced　by　accessing　information　and　human　decision-making　inevitably.　Therefore,　the　paper　proposed　an　evacuation　model　based　on　information　transfer　and　rerouting,　mainly　concentrating　on　panic　propagation,　human　decision　process　and　human　interaction,　in　order　to　examine　the　effects　of　these　factors　on　evacuation.　Results　show　that　emergency　information　trans-mission　is　crucial　in　accelerating　evacuation　since　stronger　information　intensity　brought　from　panic　propagation　improves　the　evacuation　efficiency.　The　guidance　of　human　flow　during　evacuation　can　unify　evacuation　behavior　and　the　diversion　of　the　crowd　at　the　exits　can　make　full　use　of　the　evacuation　capacity　of　exits.　Before　the　timestamp　of　39,　which　corresponds　to　the　time　at　which　the　people　reach　the　exit,　the　human　crowd　walking　in　the　corridor　generally　evacuate　faster　in　the　unified　direction　under　the　control　of　the　herd　mentality　model.　After　that,　masses　gathered　at　the　exit,　and　the　social　force　model　displays　a　better　evacuation　effect　because　of　crowd　diversion.　Spontaneous　rerouting　may　aggravate　the　evacuation　chaos,　while　scientific　evacuation　route　guidance　and　command　is　more　reasonable　and　reliable.　These　conclusions　will　guide　the　process　in　making　evacuation　plans　for　the　metro　station　hall　and　provide　a　new　perspective　to　examine　evacuation　by　considering　human　factors　that　are　closer　to　the　real　situation.