A　new　video　stitch　algorithm　based　on　dynamic　foreground　extraction　is　proposed　in　this　paper.　The　algorithm　is　found　on　static　image　mosaic.　To　solve　the　problem　caused　by　the　moving　objects　in　the　videos,　the　background　images　are　extracted.　Using　the　featured-based　method　to　get　the　transformation　matrix　between　the　background　images,　a　wide　background　video　is　stitched　up.　This　method　greatly　improves　the　efficiency　of　the　algorithm.　Besides,　the　moving　foreground　objects　are　extracted　and　the　comparability　between　objects　in　the　overlap　area　of　different　videos　is　computed　to　decide　whether　they　are　the　same　object　and　which　to　be　reserved　in　the　wide　background　video.　As　a　result,　the　ghost　is　removed.　Experimental　results　show　that　it　is　a　highly-matched　and　robust　video　stitch　algorithm,　and　can　basically　achieve　real-time　stitch.