Abstract Many studies showed that the particle shape has a significant effect on mechanical properties of sands. We are investigating the effect of proppant angularity on the dynamic conductivity of a fracture filled with proppant by simulating the irregularly shaped proppant. Our previous publication showed the important effect of grain size distribution. We showed numerically with the combination of Discrete Element Method (DEM) and Lattice Boltzmann Method (LBM) that well--graded proppant--pack can keep the fracture open for longer time. Following the previous study, we are focusing on the proppant shape effect on the conductivity of a well-graded proppant--pack. By applying sphere clumping method in DEM, we generated the semi-real shape of sand. Four groups of grains with different angularities, including high angular, medium angular, rounded grains, and spherical particles are generated randomly and packed uniformly. By applying the real condition of fracture on each proppant--pack numerically with DEM, we simulated the physics behind the fracture closing and monitored the width of the proppant--pack/fracture dynamically. Simultaneously, we calculated permeability of each proppant--pack by LBM to be able to evaluate the dynamic conductivity of each proppant--pack. Our results showed the effect of angularity on the stability of the proppant--pack. Angular proppant could keep the fracture open for the much longer time in comparison with spherical or rounded proppant grains. Furthermore, it showed that angularity can affect the permeability of the proppant--pack and well graded angular proppant--pack has lower permeability and spherical proppant-pack has the higher permeability. We reached this conclusion that there is a level in angularity of proppant that we have the most efficient pack of proppant for higher conductivity. The proposed workflow provides an effective approach for analyzing production dynamics related to variable proppant conductivity and provides a path for optimizing the proppant size distribution and shape based on expected variations in the stress field.