A comprehensive literature survey of multivariate statistical process monitoring methods of recent years is presented. Principle component analysis based methods are reviewed according to their emphases on either data attributes, such as missing value, outliers, nonlinear, time-varying, serial correlation, non-Gaussian distribution and multi-scale, or operational attributes such as multi-block, multi-mode, transition process, multi-stage. All the methods mentioned in this survey can be extended to other statistical models easily.