Whether　the　equipment　terminal　with　satellite　positioning　is　in　the　detection　method　within　the　set　range　deserves　attention,　which　determines　the　success　or　failure　of　electronic　fence　design.　The　idea　of　classification　is　now　intended　to　be　used　in　such　detection,　this　paper　proposes　an　entry　fence　detection　algorithm　based　on　support　vector　machine　(SVM).　The　algorithm　selects　the　appropriate　kernel　function　according　to　the　positioning　data　characteristics　of　the　satellite　positioning　terminal.　The　kernel　function　and　error　penalty　parameters　are　obtained　by　grid　search　method.　The　detection　accuracy　is　adjusted　to　determine　the　optimal　threshold.　The　algorithm　verification　with　the　measured　data　shows　that　the　accuracy　of　entry　fence　detection　can　reach　80-96%,　which　is　much　higher　than　the　accuracy　of　43-67%　directly　obtained,　indicating　that　the　algorithm　is　reliable　and　effective　and　has　high　practical　value.　©　2019　IEEE.