Ongoing climate change directly affects tree physiology and forest productivity but is also likely to favor an amplification of diseases and parasitism, thus increasing the global vulnerability of forest ecosystems. Tree phenology reflects the adaptive responses of forest to climate change. Phenological trends studies based on remote sensing data appear to be an efficient and costeffective way to monitor forest condition over time. Although such vegetation trends have been widely investigated at the global scale still little is known about phenophase anomalies at more finer scales (from the landscape to regional scale) and at the tree species level. Moreover, computes phenological indicators (i.e. start, end, length of the season) or disturbances indices still need to be validated and related to ecological processes. Sufficient longterm satellite data time series at coarse spatial resolution (≤ 1km) are now available to fill this gap on a broader range of spatial scale. Our retrospective analysis is based on these longterm satellite data, using the SPOT VEGETATION and MODIS sensors over the whole French forests and for the last decade. We evaluated and compared the use of several biophysical parameters (NDVI, EVI, LAI, fAPAR, fCOVER) from the latest reprocessed products (MODIS collection 5, MODIS Combined, CYCLOPES, and GEOV1) to (i) discriminate key phenological dates and (ii) quantify the annual tree defoliation status. Chronosequences of original vegetation indices were weighted, gapfilled and filtered using the adaptive StavinskyGolay filter of the TIMESAT software. Annual phenological and sanitary indicators were extracted and validated thanks to 10 years of ground observations provided by french permanent plot networks for forest monitoring (ICPForest n=540 stands and RENECOFOR network n=102 stands). The spatial responses of the main broadleaved and coniferous tree species level was investigated thanks to the systematic plot network of the National Forest Inventory (>33000 stands). The spatial patterns of phenophases and their anomalies could be depicted at the tree species level using remotelysensed data time series at coarse spatial resolution. When computing phenological dating, we found more differences between sensors than between different biophysical parameters of a same data collection. Both latitudinal and altitudinal temperature gradients could be retrieved as well as discrepancies between tree species. We also showed that tree defoliation level recorded on the ICPforest network, which is a proxy of tree health, could be efficiently monitored using LAI seasonal curves. Vegetation index interannual anomalies can also be used to investigate forest disturbances and recovery. A spatially explicit analysis revealed the consistency between the satellite- based disturbances index and large scale abiotic (i.e. drought 2003 and 2011) or biotic (i.e. forest pests) damage inventories. Based on those results, we review the abilities of longterm data series for near real time forest health monitoring and the perspectives of applications of the future satellite missions SENTINEL2 combining high temporal and spatial resolution.