A new approach is proposed to model the elastic properties of polymer nanocomposites taking into account agglomeration effects. In particular, the stiffening effect provided by rigid nanoparticles forming primary aggregates is modelled on the hypothesis that part of the polymer matrix is mechanically constrained within the aggregates. To validate the model, linear-low-density polyethylene (LLDPE)/silica micro- and nano-composites have been prepared by melt compounding followed by hot pressing. Electron microscopy observations indicated that the microstructure of the resulting nanocomposites clearly manifested primary aggregation of nanoparticles. Concurrently, thermal calorimetry and X-ray diffraction analyses proved that the crystallization behaviour was not affected by the presence of the filler. Dog-bone specimens have been mechanically tested under uniaxial tension and the data used to validate the model. A good agreement between theoretical predictions and experimental data was demonstrated. The results coupled with the propensity of nanoparticles to form aggregates could explain significant modulus increases in many nanocomposites systems reported earlier.