The paper initially uses cluster analysis to normalize and discover relationships between multi-source financial indicators, removing unnecessary data and choosing representative metrics. The extraction of common factors is performed through factor analysis, which reduces high dimensional indicators to five major dimensions. A more efficient ID3 decision tree algorithm and the Prophet time-series prediction model are subsequently incorporated to develop the financial condition early warning system. Empirical analysis relies on 16 essential financial indicators of a listed manufacturing company that existed between 2020 and 2024. The factor analysis model is 81.58percent cumulative variance contributor, and it has successfully derived five factors, which have clear economic meaning. The modified ID3 decision tree still has a general classification rate of at least 87 percent when given 50 percent category noise, and is significantly better than the traditional models. The overall score of the company in the financial field increased by 78.45 points or 6.42 points per year (1.68 per quarter) in the period of 2020-2024, which suggests the improvement of its financial performance throughout the years.
