Environmental DNA (eDNA) technology has emerged as a revolutionary tool in the ecology research, yet its application in phytoplankton monitoring remains relatively limited. To investigate the impact of human activity on phytoplankton communities in plateau wetlands, water samples were collected from both the north side (high-disturbed area of anthropogenic activities) and south side (low-disturbed area) of Lake Caohai, Guizhou Province. The composition, biodiversity, and environmental correlation of the phytoplankton community were analyzed by the techniques of metagenomics combined with eDNA. Results showed that the phytoplankton communities in Lake Caohai included 9 phyla, 58 families, 101 genera, and 152 species. In the taxa of phylum, Cyanophyta was dominant with an average percentage of 93.57%±3.06%. At genus level,Microcystis ,Aphanizomenon, Cyanobium , andSnowellawere dominant. While at the species level,Microcystis aeruginosaandAphanizomenon flos-aquaewere dominant. Furthermore, it was observed that areas high disturbed exhibited significantly lower biodiversity in phytoplankton communities compared to low-disturbed areas. Principal component analysis (PCA) indicated distinct differences in phytoplankton community composition between the two regions. LEfSe analysis further revealed that Microcystaceae, Synechococcaceae and Coelosphaeriaceae were the biomarkers in the high-disturbed area, while Aphanizomenonaceae, and Microcoleaceae in the low-disturbance area. Mantel test analysis, co-occurrence network analysis, and redundancy analysis (RDA) showed that total phosphorus was the key environmental factor affecting phytoplankton communities. Human activities influenced the biomass, biodiversity, and community structure of phytoplankton by changing the key water environment such as total phosphorus and total dissolved solids. In this paper, eDNA metagenomics technology was employed to elucidate the phytoplankton communities changes in Lake Caohai Wetland under different human activities, and provided valuable insights for the application of this technology in biological monitoring and phytoplankton assessment in plateau wetlands.