To　solve　the　problems　of　high　computation　and　low　robustness　that　exist　in　present　video　copy　detection　methods,　a　new　video　copy　detection　method　is　proposed.　This　algorithm　uses　spatio-temporal　feature　curves　of　the　video　to　describe　the　video　content.　Each　frame　in　the　video　is　divided　into　several　blocks　according　to　the　entropy　in　it　and　each　block＇s　spatio-temporal　feature　curves　are　extracted　to　serve　as　descriptors　of　its　content.　A　robust　matching　method　is　adopted　to　match　the　two　videos＇　feature　curves.　The　experiments　show　the　effectiveness　and　robustness　of　the　algorithm.