Abstract: The most personalized recommendation method research of online shopping faces challenges about how to ensure the validity of data during the data sharing process and protect users' personal privacy. In this paper, we propose a privacy preserving scheme of online recommendation based on the SVD algorithm by the truncated SVD update algorithms and randomization techniques. It turns out that the proposed scheme can conduct data efficiently and protect data privacy effectively. It is of important guiding significance for privacy preserving of online consumers.