This study constructs a composite indicator system for vulnerability assessment based on disaster-causing factors and hazard-bearing bodies involved in chemical safety accidents. In such a context, a hybrid model (D-FDM) is built by combining a data envelopment analysis (DEA) and fuzzy decision-making to take the exposure, sensitivity, and adaptability of chemical enterprises into account. A case example is employed to verify the hybrid model and demonstrate its practical application in a safety vulnerability assessment of an ammonia-producing plant in Sichuan, in southwestern China. The degrees of safety vulnerability related to the production and supporting facilities are discerned to provide insights into risk management and control for the case plant. Limitations related to the applicability of the methodology are given to lay the foundation for further improvement.