Abstract: A method, a computer device and a readable medium user's intent mining are disclosed. The method includes: obtaining text feature information corresponding to behavior data of a user; extracting an intent information list corresponding to the behavior data by using a pre-trained intent extracting model; and obtaining the target intent information of the user from the intent information list corresponding to the behavior data of the user, according to the text feature information and the intent information list corresponding to the behavior data, and a preset semantic similarity threshold, and adding the target intent information of the user to an intent information queue for advertisement recommendation.