Land cover changes resulting from urban sprawl, transport network intensification and agricultural changes contribute to fragment wildlife habitats and may lead to question viability of animal and plant populations. Landscape ecologists have shown that populations living in fragmented habitats are forced to adopt specific dynamics (patchy populations or metapopulations) making them highly dependent on fluxes between their habitat patches. Landscape connectivity may be defined as the functional response of a given species in terms of movement and individual fluxes to the potential links between habitat patches provided by the landscape structure. Many methods are used to study landscape connectivity, but following the recent literature, the graph-theoretic approach and circuit theory have been found relevant for landscape management on regional scale. Therefore, several network measures derived from graph theory have been adapted and developed by landscape ecologist to quantify habitat connectivity on graph structure. In this context, the modelling of ecological networks is a suitable approach for achieving diagnoses, prioritizing actions (for instance ecological restoration) and comparing prospective scenarios of development and conservation. This kind of modelling happily complement approaches carried out by field specialists in ecology. Graphab software is devoted to the modelling of ecological networks in the framework of graph theory. It is the only tool able to include construction and visualization of graphs, connectivity analyses and links with external data. It is easily compatible with Geographical Information Systems. Graphab is composed of 4 modules for: (1)constructing graphs, including loading initial landscape data and identifying patches and links (Euclidean distances or least-cost paths); (2)computing connectivity metrics from graphs; (3)integrating graph-based connectivity metrics into species distribution models; (4)visual and cartographic interfacing. Here we focus on our understanding and learning from experiences on the use of landscape graphs through a set of research studies concerning several species (Hyla arborea, Arvicola terrestris, Rhinopithecus bieti) in different landscape contexts (Franche-Comté and Auvergne region in France, and Yunnan in China). Future developments of the software and future applications of the graph modelling approach could be discussed.