The　risk　prediction　of　financial　markets　is　of　paramount　importance,　with　investor　sentiment　playing　a　critical　role.　However,　current　research　appears　to　be　lacking　in-depth　exploration　of　this　particular　aspect　within　the　Bitcoin　market.　This　study　aims　to　explore　the　impact　of　market　participants＇　sentiment　on　risk　prediction　in　the　bitcoin　market.　We　first　applied　ChatGPT　to　analyze　the　sentiment　of　crawled　Bitcoin-related　news　headlines.　Meanwhile,　Monte　Carlo　simulation　was　employed　to　calculate　value　at　risk　(VaR).　And　we　selected　five　conventional　factors,　including　Bitcoin　price,　transaction　volume,　market　share,　hash　rate,　and　average　difficulty　of　mining.　Finally,　K-Nearest　Neighbors　(KNN)　regression　model　was　used　to　construct　the　model　for　predicting　the　risk　of　bitcoin　market.　We　made　a　comparison　between　the　accuracy　outcomes　when　considering　and　not　considering　sentiment　as　factors.　The　results　show　that　market　participant＇s　sentiment　is　significantly　associated　with　market　risk,　and　the　inclusion　of　sentiment　can　significantly　improve　the　accuracy　of　the　risk　prediction　model.　©　2024　The　Authors.