This paper takes the enterprise tax risk early warning as the core objective, combines with the crisis management theory, and constructs the multidimensional tax risk early warning indicator system. Using multiple logistic regression method to analyze the significant indicators related to tax risk. By empirically analyzing the enterprise financial data, the risk prediction model is established and the model prediction effect is verified. The regression analysis shows that the asset-liability ratio, the proportion of financial personnel, and the return on net assets are significantly related to tax risk at the 95% confidence level. Capital intensity is significantly associated with tax risk at 90% confidence level. The established multivariate logistic regression model has a steeper gain curve and higher prediction accuracy. This study proposes strategies covering capital structure optimization application and other strategies through quantitative analysis.
