Habitat fragmentation is a threat to biodiversity because it restricts the ability of animals to move. Maintaining landscape connectivity could promote connections between habitat patches, which is extremely important for the preservation of gene flow and population viability. This paper aims to evaluate the landscape connectivity of forest areas as it relates to the conservation of the Yunnan snub-nosed monkey (Rhinopithecus bieti), an emblematic and endemic endangered primate species. Specifically, this study seeks to model ways to improve connectivity via cropland reforestation scenarios which incorporate graph theory and genetic distances. The connectivity improvement assessment is performed at two nested scales. At the regional scale, the aim is to quantitatively assess the gain in connectivity from different reforestation scenarios, in which croplands are replaced by different kinds of forest habitats. At the local scale, the goal is to prioritize and to locate croplands based on the gain in connectivity that they would provide if they were reforested. The results indicate that the four reforestation scenarios have different impacts on connectivity; the fourth scenario, in which reforestation is accomplished with plant species that provide optimal monkey habitat, yields the greatest increase in connectivity (+24% versus less than +2% for the others). Prioritization of the 1482 cropland patches shows that the 10 best patches increase connectivity from 0.04% to 9.1% as the isolation threshold distance increases. This kind of graph theoretic approach appears to be a useful tool for connectivity assessment and the development of conservation measures for species impacted by human activities.