In the face of massive item information, it is difficult for users to quickly select the choice that meets their personal preference, so personalized recommendation technology has become an important research direction to improve user experience. This paper proposes a personalized recommendation model that combines contextual information and collaborative filtering recommendation algorithm based on items. By introducing contextual information such as user time and location, the recommendation accuracy of traditional collaborative filtering algorithms is improved, and the applicability of the algorithm in dynamic scenarios is enhanced. This study designed and implemented the related prototype system, the platform was developed using Java language, and the system was constructed and verified by combining SpringBoot framework and MySQL database. The experimental results show that the model has obvious advantages in recommendation accuracy. In order to further verify the user satisfaction index, this study confirmed the improvement of user satisfaction through the user rating analysis method, which provided a reference for the further research and application of personalized recommendation technology.