Laser powder bed fusion (L-PBF) is one of the most promising additive manufacturing technologies for metals. In this work, the discrete element method (DEM) was used to reproduce a powder bed of particles distributed in a random way to be as close as possible to reality. Single and multiple scan tracks simulations were performed for Ti6Al4V alloy using a commercial CFD software, FLOW-3D AM®. The output from the numerical simulations was elaborated to obtain shape and size of melt pools, morphology of scan track surfaces and porosity content. In particular, a specific model was used in order to predict air entrainment in the melt pool and, therefore, to estimate gas porosity content, as an innovative approach to predict such defects. Results from simulations were compared with experimental data from Ti6Al4V samples produced by L-PBF in terms of melt pools size and morphology, as well as density. The good agreement between calculated and experimental results indicates that simulation of L-PBF can represent a powerful tool not only for the optimization of process parameters, but also for the prediction of porosity level.