Abstract: With the widespread application of artificial intelligence (AI), its combination with quantitative investment has promoted the rapid development of the field of quantitative finance. Based on the AI algorithms, various algorithms such as support vector machine, random forest, Adaboost and XGBoost are introduced into the multi-factor stock selection model in the field of quantitative investment, and different multi-factor quantitative investment strategies are constructed, and their investment in the new energy industry are used. A comparative study of stock picking performance. Select 35 factors out of the ten categories of factors closely related to the new energy industry, including valuation, scale, technology, and momentum, and screen out 5 effective factors through the IC value test method, and use the AT-edu quantitative platform conduct rolling backtests of quantitative investment strategies from January 2020 to December 2021, and systematically compare the performance of different multi-factor stock selection models. Finally, optimize the strategy and construct a quantitative investment strategy for the new energy industry that meets investors' expected interests and risk demands. The empirical results show that: no matter it is the traditional multi-factor stock selection model or the multi-factor stock selection model combined with the algorithm, the net value curve of the strategy is higher than the CSI 300 benchmark curve in the same period. Moreover, compared with the traditional multi-factor stock selection model, the performance of the multi-factor quantitative stock selection model constructed by the AI algorithm is better. Among them, the XGBoost stock selection model obtained the best return during the strategy backtest period, and has the largest annualized return rate and Sharpe ratio.