A method for acquiring and migrating website data, used for vertical e-commerce websites, migrating historical data information of existing e-commerce websites to other e-commerce websites through transfer learning, and collecting basic user information of existing e-commerce websites through scripts As well as user behavior, the user feature database is obtained from the basic user information, the product feature database is obtained from the product information of the existing e-commerce website, and the correlation between user features and product features is analyzed according to user behavior to obtain a correlation matrix. Other e-commerce According to the basic information of the user who enters the website, combined with the correlation matrix, the website calculates the correlation between the user and the products of the e-commerce website, and recommends them. The present invention learns the relevant patterns of user characteristics and product characteristics in the original system with dense historical data, and migrates them to other similar e-commerce websites that cannot be effectively recommended due to insufficient historical data based on the transfer learning method.