With the deep integration of artificial intelligence (AI) and Internet of Things (IoT) technologies, the maritime transportation industry is undergoing profound changes, and the application scenarios of unmanned surface vehicles (USVs) are constantly expanding. Aiming at the security threats faced by USV application systems, this paper proposes a new threat model for ship communication protocols and service logics, and the system covers the potential adversarial entity attack paths in application program logics and communication protocols. Based on this model, the automated security inspection framework USVScout was designed and implemented. By parsing the source code of the application program, the security analysis was formalized as an attribute verification task under the standard computing model, significantly improving the systematicness and accuracy of the detection. The experimental results show that in three real Marine application scenarios, USVScout successfully identified new types of vulnerabilities that static analysis tools failed to detect, and it can achieve sub-second real-time detection of 12 types of vulnerability patterns on an ordinary laptop, verifying the significant advantages of the method in terms of security and efficiency. The threat modeling and automated detection framework proposed in this paper provides technical support for ship network attack and defense drills and lays the foundation for building an intelligent and secure maritime transportation system.