The role of sentiment analysis in modern and contemporary Chinese literary works is very important, but there is a problem of inaccuracy of sentiment analysis. Traditional analysis methods cannot solve the problem of sentiment analysis in modern and contemporary Chinese literary works, and the analysis results are not reasonable. Therefore, this paper proposes a deep learning for sentiment analysis and analysis. First, the computer is used to analyze the content of literary works, and the indicators are divided according to the requirements of sentiment analysis to reduce sentiment analysis in the interfering factor. Then, the computer analyzes the sentiment of Chinese modern and contemporary literary works to form a sentiment analysis scheme, and conducts the sentiment analysis results Comprehensive analysis. MATLAB simulation shows that under certain evaluation criteria, deep learning has the accuracy of sentiment analysis of Chinese modern and contemporary literary works The rationality of sentiment analysis is better than that of traditional analysis methods.