Clustering strongly affects the transverse (T2) relaxation induced by superparamagnetic nanoparticles in magnetic resonance experiments. In this study, we used Monte Carlo simulations to investigate systematically the relationship between T2 values and the geometric parameters of nanoparticle clusters. We computed relaxation as a function of particle size, number of particles per cluster, interparticle distance, and cluster shape (compact vs. linear). We found that compact clusters induced relaxation equivalent to similarly sized single particles. For small particles, the shape and density of clusters had a significant effect on T2. In contrast, for larger particles, T2 relaxation was relatively independent of cluster geometry until interparticle distances within a cluster exceeded ten times the particle diameter. Results from our simulations suggest principles for the design of nanoparticle aggregation-based sensors for MRI.