How　to　accurately　predict　the　Dow　Jones　index　is　an　important　issue　in　quantitative　finance.　This　paper　designs　an　LSTM-AdaBoost(Long　Short-Term　Memory　and　Adaptive　Boosting)　prediction　model　based　on　the　Dow　Jones　index＇s　characteristics.　The　AdaBoost　algorithm　was　used　to　integrate　multiple　LSTM　weak　classifiers.　By　iteratively　updating　the　weights　of　each　version　and　combining　the　strategy,　the　improved　robust　classifier　was　formed,　and　the　prediction　for　the　Dow　Jones　index　was　finally　obtained.　Experimental　results　have　shown　that　the　prediction　results　obtained　by　the　LSTM-AdaBoost　model　have　been　significantly　higher　than　those　of　the　traditional　classification　model,　with　an　average　increase　of　43.76　percent　in　R_square.　The　establishment　of　this　model　provides　a　theoretical　basis　for　predicting　the　Dow　Jones　index　and　provides　ideas　for　the　intersection　of　finance　and　computer　industries.　©　2021　IEEE.