In　fingerprint　recognition,　the　feature　extraction　is　the　most　important　part.　This　part　lead　to　the　success　of　the　AFRS.　What　fingerprint　thinning　method　determines　what　feature　extraction.　The　mid-scan　thinning　method　is　an　algorithm　based　on　the　direction　map.　This　algorithm　is　more　quick.　The　thinning　map　has　less　burrs,　so　it　does　not　need　post-processing.　But　in　the　process　of　thinning　the　fingerprint　map,　the　feature　points　always　become　breakpoints.　Traditional　feature　extraction　algorithm　could　not　extract　the　feature　of　this　kind　of　thinning　map.　We　have　studied　a　new　feature　extraction　algorithm　based　on　the　breakpoints　of　the　fingerprint　thinning　map　and　the　direction　map.　This　method　solved　the　feature　extraction　problem　of　the　mid-scan　thinning　fingerprint　map.　The　experiment　shows　that　the　algorithm　has　got　good　effect.