Transformers are vital components of the electric power system, serving as the backbone for electric power transmission and transformation. Its outstanding safety performance ensures the reliable and stable operation of the power system. In this article, multi-frequency ultrasound is utilized for liquids with different physicochemical properties so as to obtain the physicochemical properties of the sample, and ultrasonic detection technology has non-destructive, real-time characteristics, so the application of ultrasonic detection technology in oil-immersed transformer oil monitoring has far-reaching significance. This paper mainly focuses on the oil-immersed transformer oil micro-water content prediction program for research, because transformer oil and water in the molecular structure of the differences, so ultrasonic detection of transformer oil and water leverages their differing ultrasonic absorption capacities. This difference is utilized to measure the micro-water content inside the transformer oil. The selected prediction model is the PCA-Elman model, which, after validation, was found to accurately predict micro-water content of transformer oil.