Journal of Information Security Research Hong Guo, Lifang Liu 16 4 2025 https://doi.org/10.6025/jisr/2025/16/4/163-172 https://www.dline.info/jisr/fulltext/v16n4/jisrv16n4_3.pdf The paper proposed a model to assess and predict financial risks in Chinese foreign trade listed companies. Recognizing limitations in the traditional Altman Z-Score model such as low diagnostic accuracy the authors integrate the Stochastic Optimization Algorithm (SOA) to fine tune the model's coefficients. Using financial data from 20 listed firms and a five year case study of Jiangsu Sainty Co., Ltd. (2012-2016), they demonstrate that the SOA optimized Z-Score model significantly improves prediction accuracy, achieving a 96.33% recognition rate outperforming SVM and AdaBoost algorithms. The analysis reveals key financial indicators (e.g., retained earnings, EBIT, and asset turnover) critically influence risk levels. Based on findings, the authors recommend enterprises establish robust risk control systems, optimize capital structures, improve working capital management, and enhance profitability. The study underscores the value of combining classical financial models with modern optimization techniques for more reliable, adaptive financial risk assessment in dynamic market environments.