In this study, a computational investigation on the mechanical properties of random carbon nanotube (CNT) networks is reported. For this purpose, atomistic models of 3-D CNT network structures are generated by an automated stochastic algorithm that is capable of randomly inserting CNTs into a design space while controlling several topological parameters including cross-link density, distance between cross-links, as well as length and chirality of CNTs. During the atomistic model generation process, heat welding is applied to the close contacts to form a welded CNT network with covalently-bonded junctions. Following the generation of atomistic models, molecular dynamics (MD) simulations of uniaxial tensile loading experiments are performed to investigate the basic deformation mechanisms of CNT networks as well as to obtain mechanical properties such as the Young's modulus, yield strength and ultimate strain values. The effects of cross-link density on the mechanical performance of CNT network materials are obtained by employing network specimens with different cross-link densities. (C) 2014 Elsevier B.V. All rights reserved.