Nanostructured anti-reflection coatings (ARC) are used to reduce the reflectivity of the front surface of solar cells. Computational electromagnetism helps to evaluate the spectral reflectivity of of this type of ARC using several approaches. They typically require large computational resources both in time and hardware elements (memory allocation, speed of processors, etc.). Long computational times may jeopardize optimization processes based on the iterative evaluation of a given merit function that depends on several parameters. Then, simplified analytic methods can speed up this evaluation with moderate computational resources. In this contribution we adapt an Effective Index Model (EIM) to the case of the design of an ARC made with nanoparticles (NP) embedded in a medium at the front surface of a thin-film silicon solar cell. Our approach modifies the discrete dipole approximation method to adapt it to the geometric and material properties of the NPs. The results obtained from the analytic method are compared with those evaluated through a Finite Element Method (FEM) for several cases involving variations in the size and geometry of the NP arrangement, obtaining reflectances that differ less than 10% for the worst case analyzed but bieng about 100 times faster than the FEM.