The fracture surface characterization, such as the fracture roughness, is essential to flow and heat transport through the fracture in the geothermal, oil, and gas reservoirs stimulation by hydraulic fracturing. This research presents an experimental approach to assess the correlation between the fracture roughness and mode I fracture toughness (KIc) of the granitic rock with different grain sizes, which is crucial in geothermal reservoirs. Three types of granite with the same mineralogical compositions but fine, medium, and coarse-grained structures were selected and prepared. Fracture toughness values were measured using a novel specimen called ENDB (Edge Notched Disc Bend), in which a cracked disc shape specimen was loaded in three-point bending. The photogrammetric method was used to estimate the value and distribution of surface roughness after ENDB testing and investigate the relationship between fracture morphology and fracture toughness. Based on the experiments, a clear correlation between the roughness and fracture toughness was observed in the three investigated granite types with different grain sizes. The FPZ around the crack tip, tortuosity, intergranular, and transgranular cracks were examined to study the effect of grain size on fracture toughness and roughness precisely. The results show that by increasing the grain size, KIc and surface roughness increased. The results also showed that the roughness parameter for the three types of rocks in the direction of the crack growth path is greater than the direction perpendicular to the crack growth path. ●The effect of grain size on the fracture toughness and roughness was considered.●Three types of granite with identical mineralogical compositions but varying grain sizes were selected.●The ENBD sample was utilized to conduct fracture toughness testing under mode I loading.●The photogrammetric method was employed to measure the roughness of the crack surface, and MATLAB code was developed to calculate surface roughness parameters.●FPZ was estimated based on thin section analysis to adequately describe the effect of grain size.