In Web 3.0 times, Internet and Electronic Commerce develop rapidly, it is necessary to solve the problem which is how to recommend the personalized information to the user when the user faces numerous of information. But now, it is only studied from three aspects: collaborative filtering, content analysis, associated rules, which belong to the two sides based on the user and the goods. All of them dig the information from the individual records of history, or the user who similar to the record of history. After analyzing the content of web 3.0, this paper point out how to mining information from the perspective of semantic-formal concept analysis based on the user, and then draw the personalized information recommendation model. After analyzing the user’s information behavior, we can find the the user’s preferences, finally recommend the proper information about commodity to the user and improve the user's satisfaction.