The development of cloud computing for processing remote sensing big data makes it possible to smoothly reconstruct global daily NDVI (normalized difference vegetation index) data. In this paper, using the Google Earth Engine (GEE) platform and based on the weighted Whittaker with dynamic λ (wWHd) vegetation index smoothing method, we propose a method of denoising daily NDVI data by interpolation before smoothing. Finally, we use the technique to obtain global smoothed daily NDVI time series using MOD09GA data.