Nowadays, problems such as low efficiency and low precision are prone to occur in the process of generating financial statements. Therefore, this paper introduces the automatic generation technology based on artificial intelligence (AI) to improve the efficiency of financial data processing and ensure the accuracy and timeliness of the statements. First, this study constructs an automatic generation system of financial statements based on artificial intelligence, adopts data cleaning and preprocessing technology to ensure the standardization and normalization of data from different financial systems, uses interpolation or mean filling method to process missing values and outliers, and converts various types of financial data into a unified format. Then, regression analysis and cluster analysis models are applied to analyze the processed data. Regression analysis is used to predict financial indicators related to historical data (such as future income, expenditure, etc.). Next, the LSTM (Long Short-Term Memory) model is used to model time series data, predict future financial trends, and provide support for the generation of accurate financial statements. Finally, combined with NLG (Natural Language Generation) technology, the Transformer model is used to automatically generate text content that conforms to the format of financial statements to ensure the format of the report is standardized and the statements are fluent. The revenue forecast accuracy for Q1–2023 reaches 91.2%, and the expenditure forecast accuracy is as high as 92.5%. These results show that the regression model can accurately capture the linear relationship between income and expenditure and effectively predict future financial trends. The automatic generation system of financial statements based on artificial intelligence effectively solves the problems existing in the traditional manual report generation process, which not only improves the efficiency of financial data processing but also ensures the accuracy and timeliness of the report content.