The hot spot temperature is an important factor affecting the operation state and insulation life of oil-immersed transformer. It is of great value to carry out multi-physical field coupling simulation research on magnetic-fluid-thermal field of oil-immersed transformer and accurately calculate and predict the hot spot temperature of transformer for transformer service life evaluation. In this paper, the oil immersed transformer is taken as the research object, and five characteristic parameters of the hysteresis model are obtained by using PSO optimization algorithm according to the experimental magnetic characteristics data of the core material and the classical J-A hysteresis model, A 3D simulation model of magnetic-fluid-thermal field is established based on the electrical and structural parameters of the oil-immersed transformer. Combined with the magnetic characteristics of the core material, the thermal field and the surrounding fluid distribution of the transformer core and winding are obtained by two-way coupling method. On this basis, in order to accurately reflect the correlation between the hot spot temperature of the transformer winding and the temperature of the oil tank wall, the selection position of the characteristic temperature point of the transformer tank wall is determined by streamline analysis method, and the hot spot temperature of the oil-immersed transformer is retrieved by support vector machine method. The results show that the prediction accuracy of the hot spot temperature reaches 0.998, and the inversion method has a high enough accuracy. It provides theoretical basis and technical support for real-time monitoring of hot spot temperature in oil-immersed transformer windings.