The wetland vegetation is an important part of coastal wetlands, and its dynamic changes affect the structures and functions of wetland ecosystem. Therefore, it is of great significance to monitor and evaluate the long-term changes of wetland vegetation by remote sensing technology for the management of coastal resources and ecological protection. In this paper, we used multi-temporal Landsat satellite images as data sources, combined object-oriented method and random forest algorithm to achieve accurate classification of typical wetland vegetation in the Current Huanghe River Estuary Reserve, and revealed the spatiotemporal variation characteristics of Phragmites australis, Suaeda salsa and Spartina alterniflora in the study area from 2000 to 2020. It has been verified that the overall accuracy of wetland vegetation mapping is between 84.74% and 92.39%, and the Kappa coefficient is between 0.81 and 0.91. The results of long time series classification show that Phragmites australis is the dominant species in the Current Huanghe Estuary Reserve, and its distribution area is maintained at more than 6% and the overall growth is steady. The area of Suaeda salsa shows a decreasing trend since 2006. The dominance degree of Suaeda salsa is decreasing continuously and the degree of fragmentation is severe. The area of Spartina alterniflora increased year by year from 211.85 hm² in 2002 to 5267.79 hm² in 2020. The expansion process of Spartina alterniflora in the reserve could be divided into three stages: in the initial expansion period before 2008, the growth of Spartina alterniflora was unstable; from 2008 to 2014, there was a rapid expansion stage, with an average annual expansion rate of 54%, which showed that the seaward expansion invaded the plain and the landward expansion invaded the habitat of Suaeda salsa in space; since 2014, Spartina alterniflora has been growing slowly, entering a stable growth period, and the annual average expansion rate is only 9%.