Evaluating the impact of moisture is critical for the production and operation of transformers. However, due to the strong magnetic fields inside transformers, it is challenging to install sensors, making it difficult to obtain information on the internal moisture distribution of the windings. This poses significant challenges to constructing a digital twin of the transformer. In response, this paper investigates a virtual-real perception method for calculating the internal moisture distribution within transformers. By integrating the real-time monitoring of load curves and moisture content in oil through physical sensors, along with the moisture balance equation, the internal moisture distribution of the transformer is obtained. The results indicate that the moisture distribution in the insulation paperboard is inversely related to the temperature distribution, with moisture concentrating at the bottom of the windings and in the laminated sheets, reaching approximately 3.2% in the 48th year.