With　the　development　of　AI　and　industrial　technology,　indoor　positioning　has　gradually　attracted　wide　attention.　General　solution　in　the　industrial　positioning　field　has　not　yet　be　found　with　current　positioning　technology.　The　problem　of　industrial　intelligent　positioning　technology　is　worth　studying.　Due　to　the　extensive　deployment　of　Wi-Fi,　this　paper　will　apply　the　improved　intelligent　fingerprint　positioning　technology　to　the　field　of　robot　positioning.　At　present,　most　fingerprint　positioning　methods　need　to　collect　a　large　amount　of　data,　and　field　investigation　requires　a　lot　of　time　and　manpower,　which　complicates　the　positioning　method.　This　paper　puts　forward　the　idea　of　synchronization　while　collecting　and　locating.　By　collecting　wifi-list　and　processed　the　information　into　fingerprint　information,　the　fingerprint　database　is　updated　regularly　in　the　meanwhile　to　eliminate　fingerprint　noise　and　reduce　the　dependence　of　location　on　environment.　The　improved　Gauss　filtering　method　is　used　to　test　the　collected　fingerprint　data　to　increase　the　reliability　of　fingerprint　data.　An　effective　matching　method　TNN　is　proposed　which　reduce　redundant　information　in　fingerprint　database,　improve　the　speed　and　accuracy　of　location　by　adjusting　fingerprint　information.　After　testing,　the　accuracy　of　the　fingerprint　positioning　method　proposed　in　this　study　reaches　2　meters,　and　the　average　accuracy　of　fingerprint　data　in　the　accuracy　range　is　98.5%,　and　what＇s　more,　it　is　easy　to　deploy　and　implement.　©　2019　IEEE.