Normalized Difference Vegetation Index (NDVI) data have been widely used for a lot of applications. Existing NDVI data services are limited by spatial and temporal scales. Using tile-based MODIS datasets in HDF format, this study has derived global NDVI datasets and implemented Web Map Service (WMS) and Web Coverage Service (WCS) (http://gis.csiss.gmu.edu/GADMFS/) to provide on-demand visualization and acquisition of global NDVI data. The web-based on-demand NDVI data services currently provide seamless 16-day composite NDVI datasets covering the globe and spanning from 2000 to 2012 without any missing data. This overcomes the spatio-temporal limitations of other existing NDVI data services. By applying advanced computing and geospatial technology, the speed of the on-demand services has been highly optimized. The software system is a contributing component of Global Earth Observation System of systems (GEOSS).