The　practical　implementation　of　hazard　analysis　and　critical　control　point　(HACCP),　and　in　particular　the　definition　of　the　critical　control　points　in　the　food　industry,　is　usually　a　complex　structured　task.　A　support　vector　machine　has　been　constructed　to　solve　this　problem.　The　proposed　methodology　allows　one　to　implement　the　first　and　second　principles　in　the　application　of　HACCP　automatically　and　intellectually.　The　CCP　identification　is　regarded　as　a　classification　problem-to　classify　the　hazards　into　CCPs　and　non-CCPs.　The　system　is　tested　through　real　cases　and　the　results　are　given.