Prosody　is　the　property　of　natural　language.　Sign　language　is　a　kind　of　language　for　the　hearing　impaired,　so　the　prosody　is　definitely　imbedded　in　sign　language.　This　makes　the　prosodic　sign　language　synthesis　to　be　necessary.　However,　the　current　research　on　sign　language　prosody　is　mainly　concerned　with　sign　language　linguistics.　This　paper　studies　the　prosody　of　Chinese　Sign　Language　(CSL)　from　the　point　of　view　of　computation.　We　first　propose　a　hierarchical　representation　model　of　CSL　prosody,　which　provides　an　initial　computational　architecture　of　CSL　synthesis　with　prosody.　Then　we　present　the　synthesis　of　CSL　＂question＂　tone　prosody　based　on　the　aforementioned　architecture.　Finally,　experimental　results　and　evaluation　show　the　enhanced　realism　and　intelligibility　of　CSL　synthesis　with　prosody.