Landscape connectivity, defined as the degree to which the landscape facilitates or impedes the movement of organisms or ecological processes among patches, has been widely recognized as one of the key objectives in studies of forest landscape management, ecological conservation and construction. The dispersal ability of organisms across changing landscapes is critical for long-term biodiversity conservation, and successful dispersal of organisms always depends on the landscape connectivity. The landscape fragmentation caused by roads has significant effects on the landscape pattern and ecological processes, such as the migration of species and biodiversity maintenance in different ecosystems of landscape. This paper aims to analyse the road impact on forest landscape connectivity in mountain area of Gongyi City, Henan Province, China. The landscape connectivity is evaluated based on a graph-theory framework, in which a graph represents a landscape as a set of nodes (habitat patches) connected to some degree by edges that join pairs of nodes functionally. Both the graph structure and the habitat availability metrics provide significant improvements for analyzing landscape connectivity and decision making in conservation planning of different ecosystems or landscapes. The geographic data sets for the study area, such as DEM, landuse, soil, vegetation and hydrology, were collected and managed in GIS. Two connectivity indicators, the probability index of connectivity (PC) and the importance value (Percentage of the variation in PC, dPC_k), were calculated. Considering the dispersal distances for different organisms, we examined the variation of landscape connectivity of the woodland in the low and middle mountainous areas according to five distance thresholds: 0.5 km, 1 km, 2.5 km, 5 km, and 10 km. The importance values of the woodland patches were calculated in the scenarios with roads and without roads in different dispersal distances. In addition, the impact of roads on the landscape connectivity of five selected woodland patches was elucidated in detail. The results shows that: (1) the woodland landscape PC values tend to increase with the distance thresholds from 0.5 km to 10 km, in both scenarios of with roads or without roads; (2) there are few patches with larger area, and in the "very high" and "high" categories of contribution to woodland landscape connectivity; (3) the dPC_k values of woodland landscape patches gradually decrease with the segmentation of roads, especially for some small patches. We conclude that the roads have important effects on the forest connectivity in the study area. The landscape indexes (such as distance threshold, node position of forest patch and forest area) and the topographical factors (elevation and slope) should be comprehensively considered when analyzing the relationships between roads and forest connectivity. Moreover, the case study also proves that the analysis of landscape connectivity could and should be used as a criterion for selecting important patches in woodland restoration planning. The analytical methods used in this paper are relatively easy to implement, and thus have application potentials in ecological restoration management and landscape planning.