With the rapid development of offshore wind power and photovoltaic power generation, the importance of marine power transmission projects is increasing and the application of submarine cables is also becoming increasingly widespread. Accurate identification of the location of submarine cables is not only beneficial for routine inspections but also enhances the efficiency of fault detection. Therefore, the routing, positioning, and fault detection of submarine cables will play a crucial role in future repair and maintenance operations. The small diameter of submarine cables and the variability of internal currents have led to a decrease in positioning accuracy and an increase in positioning difficulty. To address the aforementioned issues, this study first derives the approximate equation for the external magnetic field of submarine cables based on the underwater environment and underwater robotics, utilizing the cable structure of three-core armored submarine cables. The underwater robotic system performs attitude adjustments based on the detected magnetic induction intensity values. Building upon this, a submarine cable positioning algorithm is proposed using an Improved Grey Wolf Optimization (IGWO) algorithm. The algorithm utilizes a fitness function based on the magnetic flux density model and incorporates an online path localization method specifically designed for submarine cable detection. Finally, the accuracy and effectiveness of the IGWO algorithm are validated through simulation experiments.