In this study, a comprehensive model was proposed in order to predict the tensile strength of nanocomposites considering the effects of the random orientation of nanoparticles, interphase properties and also the inevitable aggregation/agglomeration phenomenon. The model was structured based on the nanoparticle shape (e. g. platelet, cylindrical and spherical) so it could perfectly demonstrate the random orientation. It was revealed that neglecting the aggregation/agglomeration of the nanoparticles ( ) drastically rises the prediction error to about 12% while considering decreased the error to about 7-9% for samples with low content of nanoparticles (< 0.5 Vol.). Moreover, the model was completely capable of characterizing the polymer/particle interphase regardless of the polymer type and the nanoparticles shape. Model verification was accomplished by comparing its predictions with the tensile test results of the prepared nanocomposite samples of different compositions (Polystyrene/Silica, Polystyrene/Graphene Oxide and Polystyrene/Carbon nano-tubes). Also, the samples were subjected to TEM in order to qualitatively evaluate the behavior of the different shaped nanoparticles in the polystyrene matrix.