In　order　to　get　a　kind　of　feasible　designing　scheme　to　improve　the　realistic　effects,　the　main　research　method　is　discussed.　Based　on　that,　two　key　questions　are　proposed:　One　of　them　is　constructing　the　visual　speech　representation　model,　the　other,　audio/visual　mapping　model.　After　analyzing　the　main　resolution　scheme　both　nationally　and　internationally,　system　scheme　and　key　research　contents　are　proposed　in　the　end.