As the ability to collect profiling data in metabolomics increases substantially with the advances in Liquid Chromatography–Mass Spectrometry (LC-MS) instruments, it is urgent to develop new and powerful data analysis approaches to match the big data collected and to extract as much meaningful information as possible from tens of thousands of molecular features. Here, we applied weighted gene co-expression network analysis (WGCNA), an algorithm popularly used in microarray or RNA sequencing, to plasma metabolomic data and demonstrated several advantages of WGCNA over conventional statistical approaches such as principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA). By using WGCNA, a large number of molecular features were clustered into a few modules to reduce the dimension of a dataset, the impact of phenotypic traits such as diet type and genotype on the plasma metabolome was evaluated quantitatively, and hub metabolites were found based on the network graph. Our work revealed that WGCNA is a very powerful tool to decipher, interpret, and visualize metabolomic datasets.