Normalized Difference Vegetation Index (NDVI) contains some important data for providing vegetation cover information and supporting environmental analyses. However, understanding long-term vegetation cover dynamics remains challenging due to the limitations of temporal and/or spatial resolution of remote sensing data. In this paper, we fused the Moderate Resolution Imaging Spectroradiometer (MODIS) data and Landsat series satellite data based on Seasonal Auto Regressive Integrated Moving Average model with Covariates (SARIMAX) for reconstructing the NDVI of the past 20 years. In a comparison with other time series analysis models, SARIMAX has the advantages of higher R² and better reconstruction capabilities.