The invention provides a power file automatic classification method and device. The method comprises: generating a corpus set, a title set and a vocabulary set according to the power files to be classified; generating a K-dimensional input vector based on the vocabulary set, wherein K is the number of vocabularies in the vocabulary set; training the input vector by using the corpus set and the title set and based on a word embedding mode, so as to enable the K-dimensional input vector to be compressed into a C-dimensional word vector; counting the word frequency of each keyword in the vocabulary set in each to-be-classified power file and generating a vector corresponding to each to-be-classified power file; constructing a matrix according to the C-dimensional word vectors, and calculatingproducts of the matrix and vectors corresponding to the to-be-classified power files; and carrying out clustering analysis on each product result, and taking the Minkowski distance as a vector distance in clustering analysis to divide the to-be-classified power files into a preset number of categories. According to the invention, the classification efficiency and the classification accuracy can be improved.