The presence of arsenic in groundwater and other drinking water sources presents a notable public health concern. Although the utilization of iron oxide nanomaterials as arsenic adsorbents has shown promising results in batch experiments, few have succeeded in using nanomaterials in filter setups. In this study, the performance of nanomaterials, supported on sand, was first compared for arsenic adsorption by conducting continuous flow experiments. Iron oxide nanoparticles (IONPs) were prepared with different synthetic methodologies to control the degree of agglomeration. IONPs were prepared by thermal decomposition or coprecipitation and compared with commercially available IONPs. Electron microscopy was used to characterize the degree of agglomeration of the pristine materials after deposition onto the sand. The column experiments showed that IONPs that presented less agglomeration and were well dispersed over the sand had a tendency to be released during water treatment. To overcome this implementation challenge, we proposed the use of clusters of iron oxide nanoparticles (cIONPs), synthesized by a solvothermal methodology, which was explored. An isotherm experiment was also conducted to determine the arsenic adsorption capacities of the iron oxide nanomaterials. cIONPs showed higher adsorption capacities (121.4 mg/g) than the other IONPs (11.1, 6.6, and 0.6 mg/g for thermal decomposition, coprecipitation, and commercially available IONPs, respectively), without the implementation issues presented by IONPs. Our results show that the use of clusters of nanoparticles of other compositions opens up the possibilities for multiple water remediation applications.