In order to improve the scientificity and accuracy of enterprise financial crisis identification, a financial crisis early warning model based on random forest algorithm is constructed. The model parameters are optimised by collecting, pre-processing and feature engineering the financial data of A-share listed companies, combined with cross-validation and grid search. Based on the comparison of logistic regression and SVM models, Random Forest performs better in terms of accuracy, F1-score and AUC, which verifies its practical value in financial early warning. The results can provide data support and decision-making reference for enterprise risk prevention and control.