We here present an approach combining virtual and nuclear magnetic resonance (NMR)-based screening to identify low molecular weight molecules (small molecules) targeting viral RNA elements from the SARS-CoV-2 genome. Our method uses high-resolution RNA structural ensembles of the conserved 5'-terminal stem-loop 1 (SL1) and the pseudoknot (PK) of the –1 programmed frameshift element as targets for binding of small molecules derived from a virtual library of compounds. The ensemble representation of the RNA structural dynamics was found to be crucial to faithfully identify RNA binders during virtual screening. Hits proposed by screening of the virtual library were probed for their binding to the two RNA elements by ligand- and target-based NMR experiments. This combined approach increases the specificity and efficiency of drug discovery by targeting specific of the multiple RNA conformations present within the RNA ensemble structure. Our results demonstrate that the integration of virtual and experimental NMR screening efficiently identifies RNA-binding small molecules as start molecular entities to advance RNA-targeted antiviral therapies. The strategy does not only apply to SARS-CoV-2, but also provides a rapid, highly specific route to discovering therapeutics for other RNA-based pathogens, highlighting the critical role of RNA structural data in enriching drug discovery efforts.