In this work, a fast and sensitive high performance liquid chromatography coupled with multivariate analysis was utilized to assist the quality assessment of 10 bathes of Rhodiola rosea extracts (RREs) from two manufacturers.131 peaks were detected in RREs by this method using a fused-core column.Principal component analysis (PCA) and hierarchical cluster analysis (HCA) of the chromatographic data showed that the 10 batches of RREs could be well differentiated, and categorized into three groups.Such a categorization was found to be closely related to the origins of RREs.Rosavin and five unknown components were screened by partial least square-discriminant analysis (PLS-DA) as the components responsible for the quality differentiation of RREs.The observation was further confirmed by the PLS prediction models that the level of rosavin was more relevant to the multivariate chromatographic data of RREs,compared to the ones of salidroside and tyrosol, the other two ingredients which are commonly used to standardize RREs.Results of the present study not only indicated that rosavain was a rational marker to control the quality of RREs, but also demonstrated the power of HPLC-UV based metabolomics in the quality assessment of herbal extracts.