Abstract: The relationship between phytoplankton and environmental factors is an important field in marine ecology. However, relationships between phytoplankton and multiple factors are complex and hard to be explained. Thus it usually needs to use multivariate statistical analysis (MSA) to describe these relationships. Based on meta-analysis, we summarized the widely used multivariate statistical analysis methods in the phytoplankton ecology field journals and compared the use frequency, application scenarios, advantages and disadvantages of each method, including cluster analysis (CA), principal component analysis (PCA), canonical correspondence analysis (CCA), discriminant analysis(DA), factor analysis (FA), redundancy analysis (RDA), non-metric multidimensional scaling(NMDS), and analysis of similarity (ANOSIM). The results showed that CA, PCA, and CCA are the most popular methods of MSA in the recent 20 years. PCA and RDA are mainly based on linear models, which are applicable to scenarios with few species and small changes in environmental factors and species abundance fluctuations. For cases of species distribution changes or environmental gradients that are highly variable, using single-peak models usually has better performance, such as CA and CCA. In some special scenarios, the performance of the MSA method with the nonlinear assumption is better than that with the linear assumption. This implied the importance of trying multiple methods and researchers' experiences. In addition, according to the performance of these methods in the historical studies, we provide a simple index of the most popular MSA methods and their application scenarios, to help researchers, especially beginners, quickly select one or more MSA methods according to their own needs.