Silver nanoparticles (NPs) functionalized with the molecular assembler bis-acridinium dication lucigenin (LG) have been used as a chemical sensor system to detect a group of polycyclic aromatic hydrocarbon (PAH) pollutants in a multicomponent mixture by means of surface-enhanced raman scattering (SERS). The effectiveness of this system was checked for a group of PAHs with different numbers of fused benzene rings, namely anthracene, pyrene, triphenylene, benzo[c]phenanthrene, chrysene, and coronene. In order to determine the host capacity of this sensor system, the self-assembly of the LG viologen on a metallic surface has been checked by analyzing SERS intensities of PAH bands at different LG concentrations. The NP-LG-analyte affinity is derived from the analysis of PAH band intensities at different concentrations of pollutants, the adsorption isotherm of each PAH on NP-LG cavities has been studied, and the corresponding adsorption constants have been evaluated. The limit of detection at trace-level concentration is confirmed by the presence of their characteristic fingerprint vibrational bands. The SERS spectra of PAH mixtures confirm that LG viologen dication shows a higher analytical selectivity to PAHs constituted by four fused benzene rings, mainly pyrene and benzo[c]phenanthrene, in agreement with their higher affinity which is also related to their better fit into the intermolecular LG cavities. As a conclusion, SERS spectra recorded on modified NP-LG surfaces are a powerful chemical tool to detect organic pollutants.