The exceptional properties of carbon nanomaterials make them ideal reinforcements for polymers. However, the main challenges in utilizing their unique properties are their tendency to form agglomerates, their non-controlled orientation, non-homogeneous distribution and finally the change in their shape/size due to processing. All the above are the result of the nanomaterial/polymer interfacial interactions which dictate the overall performance of the composites including the mechanical properties. The aforementioned uncertainties are the reason for the deviation observed between the experimentally determined properties and the theoretically expected ones. The focus of this study is to understand the reinforcing efficiency of carbon nanomaterials in polymers through finite element modeling that captures the effect of the interfacial interactions on the tensile modulus of polymer nanocomposites (PNCs). The novelty of this work is that the probability distribution functions of nanomaterials dispersion, distribution, orientation and waviness, determined through image analysis by extracting 3-D information from 2-D scanning electron micrographs, are incorporated into the finite element model allowing thus for fundamental understanding of how the nanostructure parameters affect the tensile modulus of the PNCs. The nanocomposites are made using melt mixing followed by either injections molding or melt spinning of fibers. Polypropylene (PP) is used as the polymer and carbon nanotubes (CNT) or exfoliated graphite nanoplatelets (xGnP) are used as nanoreinforcements. The presence of interphase, confirmed and characterized in terms of stiffness and width using atomic force microscopy, is also accounted for in the model. The dispersion and distribution of CNT within the polymer is experimentally altered by using a surfactant and by forcing the molten material to flow through a narrow orifice (melt spinning) that promotes alignment of CNT and even of the polymer chains along the flow/drawing direction. The effect of nanomaterials' geometry on the mechanical behavior of PNCs is also studied by comparing the properties of CNT/PP to those of xGnP/PP composites. Finally the reinforcing efficiency of CNT is determined independently of the viscoelastic behavior of the polymer by conducting tensile testing at temperatures below the glass transition temperature of PP. The finite element model with the incorporated image analysis subroutine has sufficient resolution to distinguish among the different cases (dispersion, distribution, geometry and alignment of nanomaterials) and the predicted tensile modulus is in agreement with the experimentally determined one. In conclusion, this study provides a tool, that integrates finite element modeling and thorough experiments that enables design of polymer nanocomposites with engineered mechanical properties.