Reacting to the industrial 4.0 era, the information integration capabilities of smart manufacturing and industrial Internet of things (IIoT) have become a cornerstone of transformation in the traditional manufacturing industry. The industrial control system (ICS) is the significant communication of the factory environment, which is the main attacking target. Therefore, how to discover the potential vulnerability is the most important issue. Fuzzy testing is one of the vulnerability exploration techniques, however, the classical one is not suitable for ICS. The coverage guided is lower than that of the information technology situation. To address this challenge, the authors propose an ICS-oriented fuzzy testing (ICS-FT) technique preserving the higher coverage guided. The algorithm of dynamic link database injection is used to improve the coverage rate within the Windows system. Especially, the authors divide the zones and conduits for the testing bed based on IEC 62,443 and Purdue model. In the experiments, the proposed approach consistently exhibits substantially improved coverage rates compared to conventional fuzz testing, such as AFL and AFLFast. Moreover, ICS-FT incorporates a significantly greater number of testing inputs. This can afford a comprehensive understanding of the testing condition.