Abstract: Firstly, the metabolites in plasma were extracted to obtain organic and aqueous phases by Folch method. Several kinds of metabolites in organic and aqueous phases were detected simultaneously by ultra-performance liquid chromatography-high resolution mass spectrometry (UPLC-HRMS). Mobile phase and elution procedure were optimized. Chromatographic separation was performed on a ACE3 C18 column (150 mm×3.0 mm×3.0 μm), and the MS analysis was carried out by electrospray ionization (ESI) source in positive and negative ion modes combined with full and segmental scanning. The mass spectrometry-based molecular networking (MSMN) has been widely used, but it has not been mentioned in the analysis of lipid metabolites. In this study, the method for the analysis of metabolites in plasma by UPLC-HRMS combined with MSMN was developed. The metabolites identified by chemical standards were selected as seed compounds, and MSMN was used to screen the metabolites having similar MS/MS behavior with the seed compounds, so as to expand the qualitative range and improve the qualitative efficiency of metabolites. A total of 187 metabolites were annotated, including 80 metabolites such as amino acid, fatty acid and sugar in the aqueous phase and 107 lipid metabolites in the organic phase, among which 64 metabolites were identified by MSMN, which achieved the high throughput analysis of multiple metabolites in human plasma in this way. It provided a feasible new idea for the identification of lipid metabolites but there were still many limitations and challenges. The MSMN is suitable for the analysis of metabolites with large molecular weight, and the prerequisite is that the MS/MS of certain metabolites contains at least 3 or more common characteristic fragments. In this work, the metabolites identified by MSMN in the organic phase were mostly phosphatidylcholines (PCs) and sphingomyelins (SMs) in positive ion mode. For example, the aggregation degree of triacylglycerols (TGs) was high in the MSMN, but it was difficult to annotate successfully because of the weak characteristic of common fragments. However, in negative ion mode, it was more difficult to use MSMN to characterize the metabolites because of the small number of characteristic fragments of congener metabolites and the low mass spectral response of some fragments, which was difficult to be detected. We also tried to establish the MSMN of metabolites in the aqueous phase and obtained the molecular network diagram with good aggregation. However, the qualitative results were not satisfactory, and the metabolites in the aqueous phase were all small molecular weight metabolites such as amino acid, and the MS/MS lacked representative characteristic fragments, so MSMN was not suitable for these metabolites.