Due to the diverse and numerous types of users， it is easy to encounter inaccurate push content caused by inaccurate user feature extraction when designing personalized promotion information push methods. Therefore， a personalized promotion information push method based on dynamic user profiles is proposed. In the study， massive user behavior data is formatted to ensure accuracy and availability；multiple user behavior characteristics is captured within the framework of dynamic user profiling； detailed classification processing on the collected promotional information is conducted in order to better understand and extract the core features of the promotional information；by calculating the matching degree between promotional information characteristics and user behavior characteristics， the accuracy of push notifications is ensured. Moreover， based on the actual preferences of users for different promotional information， a personalized push function is constructed to dynamically adjust the push strategy to meet the personalized needs of users. In the experimental testing， the user conversion percentage of the proposed method is higher than 0.95， indicating high accuracy in practical applications.