The distribution of ecosystem sites, their connectivity and their evolution are key links between ecology and society, as these landscape patterns determine the regional sustainability for biodiversity and are subject to human interventions. While little attention has been paid to procedures for mapping connectivity among numerous “natural” landscape elements over large areas, the principal objective of this study was to map ecological corridors at a regional scale using remotely sensed data. To achieve this goal, we developed an approach including three stages: first, permanent “natural” landscape elements were identified from Landsat TM images; second, permanent grassland connected to these elements was extracted from a multitemporal series of MODIS images and aggregated to them; third, connectivity was depicted by least-cost modeling. The regional connectivity map can be analyzed at a regional scale to highlight large corridors, but also at a local scale to determine areas for conservation intervention.