In this article, the efficiency of all the well-known valency-based molecular descriptors has been measured using certain comparative testing. Importantly, we use a computational technique for hexagonal systems to generate regression models to determine the correlation of valency-based molecular descriptors with the π-electronic energies for polycyclic aromatic hydrocarbons. The study implicates some favorable outcomes as the sum-connectivity index and certain Randić index have correlation coefficient of 0.9999. To ensure the applicability of our proposed study, we compute analytically explicit expressions for efficient valency-based molecular descriptors for certain infinite families of carbon nanotubes and carbon nanocones. Our results assist in correlating the π-electronic energies of underlying chemical structures of these nanotubes and nanocones.