We　turn　to　statistical-based　methods　and　propose　a　statistical　inference　approach　by　using　Dynamic　Bayesian　Network,　which　is　able　to　learn　automatically　from　data　set.　By　soccer　video　analysis　is　as　an　example,　the　proposed　method　is　verified　by　experiment.　We　extract　the　color,　shape　and　other　low-level　features　from　the　video,　to　detect　and　identify　5　kinds　of　high-level　semantic　events　using　dynamic　Bayesian　network　model.　The　experimental　results　show　that　our　method　is　effective.　©　2015　SERSC.