Aiming　at　the　characteristic　of　distortion　in　the　evaluated　result　because　of　much　unascertanty　information　and　the　difficulty　to　synthesize　the　unascertanty　and　fix　quantify　to　evaluate　the　scheme　in　the　selection　of　the　long-span　structure,　the　unascertained　measure　model　and　the　incredible　degree　are　discussed　and　a　model　of　selection　of　the　long-span　structural　projects　based　on　the　unascertained　measure　is　proposed　in　this　paper.　The　model　has　more　advantage　than　fuzzy　math　when　disposing　unascertained　factors,　e.g.　less　data　lost　and　more　credible.　Finally,　an　engineering　example　is　discussed　by　using　the　proposed　model.