Urban　rail　transit　has　been　commonly　regarded　as　an　efficient　transportation　mode　for　alleviating　the　negative　impacts　of　rapid　urbanization.　However,　the　complex　and　comprehensive　built　environment　caused　by　rapid　urbanization　prevents　policymakers　and　planners　from　understanding　the　performance　of　urban　rail　transit,　particularly　around　transit　stations.　Developing　station　area　typologies　plays　an　important　role　in　describing　the　characteristics　of　station　areas.　Numerous　studies　have　developed　typologies　of　station　areas　over　the　past　decades;　however,　few　have　attempted　to　estimate　the　distribution　of　station　types　per　station.　To　fill　this　gap,　this　paper　first　proposed　the　importance　of　a　land　use　type　(ILUT)　index　based　on　the　term　frequency-inverse　document　frequency　method.　This　method　avoids　mutual　concealment　among　land　use　types,　to　extract　key　built　environment　features　based　on　built　area　datasets　of　the　buffer　areas　of　337　selected　urban　rail　stations　in　Beijing.　Subsequently,　the　ILUT　index　was　used　to　estimate　the　distribution　of　the　station　types　per　station　using　the　latent　Dirichlet　allocation　model.　The　12　identified　types　of　urban　rail　station　areas　in　Beijing　demonstrated　that　land　use　types　consist　of　station　types　and　presented　the　distribution　of　the　station　types　with　their　probabilities,　revealing　the　current　performance　of　station　areas.　This　paper　proposes　a　description　approach　that　will　help　policymakers　and　planners　to　effectively　and　efficiently　identify　and　analyze　built　environment　features.