Temperature and ampacity are two crucial parameters to characterize the operation and thermal stability conditions of a power cable. The value of the actual soil thermal conductivity is essential for the accurate calculation of the cable temperature field. However, the soil thermal conductivity exhibits significant variations and uncertainties due to different factors. To accurately calculate the temperature field and ampacity, a mathematical model and the corresponding solution methodology are proposed to inverse the real-time soil thermal conductivity. Firstly, a distributed optical fiber temperature measurement system (DTS) is developed to obtain the temperature of the cable sheath. Secondly, based on the real-time temperature from DTS, a mathematical model of real-time soil thermal conductivity inversion under the cable duct laying is constructed. Finally, a particle swarm optimization algorithm (PSO) for solving the corresponding inverse problem is proposed. The numerical results on a case study indicate that the real-time ampacity obtained by using the inversed real-time soil thermal conductivity from the proposed methodology is approximately 41.85% higher as compared to the nominal ampacity.