Enterprise bankruptcy prediction is essential for financial investment and corporate management. Traditional machine learning based methods mainly focus on internal risk mining while ignoring the external risks propagated among enterprises. Recent research efforts have attempted to use graph neural networks to model external risks. However, these methods generally lack interpretability and suffer from transferability. To address these issues, we propose a novel enterprise bankruptcy prediction model with meta-path denoising and capsule network modeling. Specifically, we design an automatic meta-path generation and selection method to improve the model’s transferability while minimizing the information loss and noise introduction. Furthermore, a hierarchical meta-path information aggregation method is designed to enhance the model’s interpretability, and a capsule network based risk assessment module is designed to dynamically capture enterprises’ risky and non-risky factors for better bankruptcy prediction. Extensive experiments demonstrate the effectiveness and interpretability of the proposed model.