Gait　recognition　is　an　important　biometric　technology,　and　also　an　important　research　direction　in　the　field　of　human-computer　interaction.　Since　the　gait　information　provided　by　each　person＇s　walking　is　different,　different　people　can　be　identified　according　to　different　gait　information.　Based　on　WiFi　CSI　analysis,　we　propose　another　gait　recognition　model　Wi-IP　(WiFi-Identity　Perception),　which　achieves　good　results　with　a　small　number　of　training　samples.　In　the　corridor　environment,　we　conducted　a　lot　of　experiments　to　ensure　the　accuracy　of　identity　recognition.　In　the　system　model　we　designed,　when　each　group　of　2　to　6　volunteers　walked　3.5　meters,　the　average　recognition　accuracy　of　our　system　was　between　82.5%　and　100%.　©　2022　IEEE.