The invention discloses an automatic corpus acquisition method and device, computer equipment and a storage medium, and relates to an artificial intelligence technology, and the method comprises the steps: training an LDA model according to each document which is marked with a topic probability distribution result in advance in a corpus to obtain the LDA model; then, using the LDA model to perform operation according to description text data of the input object to obtain a predicted topic probability distribution result, and finally, according to the predicted topic probability distribution result, selecting corpora with the same topic from a corpus to form a first target corpus subset, and sending the first target corpus subset to a user side. Through the LDA model, the corpus required in the interview process is predicted, and the target corpus is automatically screened and pushed based on the predicted prediction subject, so that the acquisition efficiency of the target text is improved, the target text is pushed more timely, and the real-time performance is improved.