Air pollution poses significant threats to public health, ecosystems, and economic stability, particularly in rapidly industrializing regions like China. This study aims to accurately predict air pollution levels emitted by Chinese publicly listed companies from 2007 to 2022 using the Gradient Boosting Regression Tree (GBRT) algorithm. By analyzing detailed environmental reports and emissions disclosures, we construct a robust predictive model to forecast emissions of Sulfur Dioxide (SO2), Nitrogen Oxides (NOx), and Dust. The model's performance, evaluated through metrics such as Mean Squared Error (MSE) and R2 score, demonstrates high accuracy and reliability, with R2 scores of 0.8457 for SO2, 0.8653 for NOx, and 0.8758 for Dust, and corresponding MSE values of 20240.71, 36928.23, and 81928.29, respectively. This predictive capability is crucial for policymakers and industry stakeholders to develop effective mitigation strategies, ensuring compliance with environmental standards and promoting corporate sustainability.