The singular value decomposition (SVD) is a widely used matrix factorization tool which underlies many useful applications, e.g. recommendation system, abnormal detection and data compression. Under the environment of emerging Internet of Things (IoT), there would be an increasing demand for data analysis. Moreover, due to the large scope of IoT, most of the data analysis work should be handled by fog computing. However, the fog computing devices may not be trustable while the data privacy is the significant concern of the users. Thus, the data privacy should be preserved when performing SVD for data analysis. In this paper, we propose a privacy-preserving fog computing framework for SVD computation. The security and performance analysis shows the practicability of the proposed framework. One application of recommendation system is introduced to show the functionality of the proposed framework.