For the fuzziness and randomness of quantitative information assessment in transformer state assessment, a normal cloud model is used for transformer state assessment. The multi-dimensional cloud algorithm used in this paper can effectively avoid the problem that it is difficult to balance the weights among various indicators. First of all, through the normal cloud expansion of the transformer operation data, cloud droplets are continuously generated using the normal cloud generator. Secondly, according to the scope of the given range of indicators to expand to the rating scale cloud, cloud droplets generated by the data cloud as input, using a multi-dimensional cloud algorithm to calculate the operating status of the transformer. In this paper, the operation information of a certain transformer is taken as the analysis object. The validity of the method is verified by an example, and the method is proved to be simple, efficient and reliable.