Water inrush disasters in karst tunnels pose a significant threat to construction safety, project timelines, and the long-term sustainability of infrastructure. Effective risk assessment is crucial for mitigating these hazards and ensuring the resilient development of underground transportation networks. This study proposes a quantitative risk assessment model that integrates the entropy weight method with attribute interval recognition theory to address the uncertainties inherent in complex geological environments. First, a hierarchical evaluation index system is established based on four primary controlling factors: stratigraphy, geological structure, topography, and hydrogeology. Subsequently, the entropy weight method is employed to objectively determine the weight of each index, thereby minimizing human bias. The attribute interval recognition model is applied to calculate the comprehensive attribute measure for each tunnel segment, effectively managing the fuzziness of risk classification boundaries. The risk grade is ultimately determined using the confidence criterion. The proposed model is applied to the Qigan Mountain karst tunnel in Chongqing, China, which is divided into 56 segments for detailed analysis. Results show 41.32% (4088 m) high-risk, 40.14% (3971 m) medium-risk and 18.55% (1835 m) low-risk sections, which are highly consistent with the theoretical water inflow calculation results. The model realizes accurate and quantitative water inrush risk assessment, providing a scientific basis for disaster prevention and control in karst tunnel construction, and further promoting the sustainability and safety of underground engineering in karst areas.