With　the　popularity　of　mobile　Internet,　Android　has　become　the　mobile　operating　system　with　the　largest　market　share,　Android　applications　have　become　a　part　of　people＇s　life,　but　the　credibility　of　Android　applications　is　increasingly　difficult　to　guarantee.　According　to　statistics,　the　number　of　Android　applications　in　the　country　has　exceeded　3　million,　in　such　a　huge　scale,　some　malicious　applications　mixed　in,　stealing　privacy,　threatening,　malicious　deduction　and　other　untrustworthy　behavior　has　caused　great　impact　on　the　life　of　users.　Most　of　the　current　research　on　Android　malicious　applications　only　focuses　on　how　to　detect　malicious　applications,　but　not　on　their　behavior.　Therefore,　in　this　paper,　we　will　take　Android　application　as　the　research　generalization,　generate　its　sensitive　API　call　graph　by　extracting　the　sensitive　API　and　permission　call　situation　of　Android　application,　extract　the　untrustworthy　behavior　of　Android　application　by　analyzing　and　studying　the　API　call　graph,　and　finally　build　a　template　of　untrustworthy　behavior　statement　for　Android　application.　The　application　market　can　detect　whether　there　is　untrustworthy　behavior　of　Android　applications　based　on　this　behavior　statement　template.　©　2022　IEEE.