In the era of artificial intelligence technology, it has become a trend to combine artificial intelligence technology with other industries. As an emerging force in artificial intelligence, ChatGPT provides new ideas for the combination of the financial industry and artificial intelligence. This paper focuses on whether it is feasible to apply ChatGPT to analyze text sentiment in financial markets. We collected three types of financial texts from different information sources, designed two rounds of experiments based on the subjectivity of the texts, “Industry Emotional Training” and “Enterprise Emotional Training”, and compared the performance of ChatGPT3.5 and ChatGPT4.0 horizontally. Experimental results show that it is feasible to use the ChatGPT tool to analyze text sentiment in financial markets. Based on the differences in the performance of different versions of ChatGPT, ChatGPT4.0 is better than ChatGPT3.5 in analyzing financial market text sentiment. These findings provide support for the development of ChatGPT in the financial industry. This paper further explores how ChatGPT can be applied to a wider range of financial scenarios in the future, allowing AI technology to better assist the digital transformation of the traditional financial industry.