Pedestrian　safety　is　one　of　the　research　focuses　all　over　the　world.　Intelligent　decision-making　makes　it　possible　to　provide　dangerous　risk　prediction.　This　paper　aims　to　serve　as　a　stepping　stone　for　avoiding　serious　fatal　vehicle　-　pedestrian　crash.　It　provides　a　method　for　intelligent　vehicles　to　identify　the　factors.　Business　and　education　Point　of　Information　(POI)　data　in　Beijing　were　collected　and　processed　to　partition　traffic　zones　into　high　economic　zones　and　low　economic　zones　used　the　method　of　k-means　clustering　algorithm.　Then　a　binary　logistic　regression　was　utilized　for　recognition　of　contributing　factors.　The　result　takes　several　important　factors　into　account　in　low　economic　zones　needed　special　attention,　such　as　fourth　class　road　and　general　city　road.　As　a　result,　the　findings　of　this　study　could　assist　to　design　the　hardware　module　and　programming　of　intelligent　vehicle　to　enable　pedestrian　safety　be　improved　over　the　long　term.