The hot-spot is the crux point inside transformers, and its temperature rise is a key parameter limiting the transformer capacity. For accurately and quickly acquiring the hot-spot temperature rise, this paper presented a inversion means for calculating hot-spot temperature rise by outer wall temperature rise of transformer tank. Firstly, a 10 kV oil-immersed transformer “fluid-temperature” multi-field simulation model was established. Secondly, the heat flow field distribution law inside the transformer was simulated and analyzed. Then, the oil path was analyzed. On this basis, the characteristic temperature measuring points located on the outer wall of the transformer tank which have a strong relevance with the hot-spot were chose. Finally, neural network algorithm was used to construct the correlation between load rate, characteristic point temperature rises and the hot-spot temperature rise. In this way, this paper realized the inversion calculation of distribution transformer hot-spot temperature rise. The relative root-mean-square error of the inversion estimated method proposed is 2.22%. This method can accurately and fastly obtain the distribution transformer hot-spot temperature rise.