Abstract: Invasive species Spartina alterniflora poses a severe threat to the ecosystems and biodiversity in coastal wetlands across China. Multiple rounds of Spartina alterniflora removal and restoration engineering have been carried out in coastal areas of Jiangsu Province. To identify the removal dynamics and monitor the recurrence situations in a timely manner, this study investigated the Jiangsu coastal wetlands as a case study. Accordingly, a removal-recurrence dynamic monitoring framework was developed based on multi-source time-series remote sensing data. First, the dense time-series curves of the normalized difference vegetation index (NDVI) were plotted by integrating Sentinel-2 and Landsat8 imagery on the Google Earth Engine (GEE) platform and combining an extreme gradient boosting (XGBoost) model. Then, the removal activities of Spartina alterniflora from 2019 to 2024 were identified using a difference-based approach, achieving an overall coefficient of determination (R²) for fitting reaching up to 0.85. Furthermore, a Siamese neural network model based on multi-scale feature extraction was developed to detect the recurrence of Spartina alterniflora, with the overall accuracy of up to 94%. The results indicate that in Yancheng City, the removal area of Spartina alterniflora increased from 35.42 km² in 2023 to 125.47 km² in 2024. The removal activities in the Lianyungang and Nantong cities were concentrated in 2023 (9.86 km²) and 2024 (38.01 km²), respectively. No recurrence was observed in the southern buffer zone of the Yancheng National Nature Reserve for Rare Birds and along the Dafeng Port in 2024. However, re-invasions occurred near the Chuandong Port and the Fangtang Estuary, covering areas of 1.12 km² and 1.29 km², respectively. The removal-recurrence dynamic monitoring framework developed in this study can provide technical support and managerial reference for assessing the effects of Spartina alterniflora control and ecological restoration in coastal wetlands.