ESG data refers to the relevant data of environmental, social, and governance aspects of enterprises, which contains a lot of information that has an impact on the sustainable development of enterprises. Through mining and analyzing ESG data, information support can be provided for the sustainable development of enterprises. This article adopted a keyword based text classification method, preprocessed ESG data about enterprises, classified them based on keywords, and trained a naive Bayesian model with these keywords. Finally, the highest classification recall rate reached over 90%. The results showed that the text classification model could effectively classify texts in ESG text data. At the same time, based on the classification results of the model, a three-dimensional radar chart was drawn from a score table of ESG data in the development of the enterprise, in order to understand the problems existing in the enterprise itself. From the results of this experiment, it can be seen that in the field of ESG data analysis for enterprises, the combination of data mining and naive Bayesian algorithms can provide favorable support for the sustainable development of enterprises.