The　rapid　growth　in　demand　for　location-based　services　(LBS)　is　mainly　due　to　the　development　of　mobile　devices　and　related　services　in　recent　years.　As　one　of　the　LBS,　the　indoor　positioning　system　(IPS)　has　become　a　research　hotspot.　Bluetooth　technology　has　become　one　of　the　preferred　indoor　localization　technologies　due　to　its　low　cost　and　convenient　installation.　At　present,　most　Bluetooth　Low　Energy　(BLE)-based　localization　methods　use　fingerprinting　approach.　However,　in　the　presence　of　large-scale　BLE　sensors,　the　traditional　fingerprinting　approach　will　take　up　a　lot　of　computing　resources　or　even　can　not　be　calculated.　Meanwhile,　the　heterogeneity　of　mobile　devices　will　also　affect　the　localization　accuracy.　In　this　paper,　we　designed　a　dynamic　selector　to　solve　the　problems　caused　by　large-scale　BLE　sensors,　and　a　dynamic　weighted　nearest　neighbor　(DWNN)　indoor　localization　algorithm　was　proposed.　Experiments　show　that　the　indoor　localization　method　proposed　in　this　paper　designs　a　dynamic　selector　to　solve　the　problem　of　large-scale　BLE　sensors.　In　addition,　the　device　heterogeneity　has　been　solved　and　higher　localization　accuracy　has　been　achieved.　©　2021　Technical　Committee　on　Control　Theory,　Chinese　Association　of　Automation.