The applicability of DFT-based descriptors for the development of toxicological structure-activity relationships is assessed. Emphasis in the present study is on the quality of DFT-based descriptors for the development of toxicological QSARs and, more specifically, on the potential of the electrophilicity concept in predicting toxicity of benzidine derivatives and the series of polyaromatic hydrocarbons (PAH) expressed in terms of their biological activity data (pIC50). First, two benzidine derivatives, which act as electron-donating agents in their interactions with biomolecules are considered. Overall toxicity in general and the most probable site of reactivity in particular are effectively described by the global and local electrophilicity parameters respectively. Interaction of two benzidine derivatives with nucleic acid (NA) bases/selected base pairs is determined using Parr’s charge transfer formula. The experimental biological activity data (pIC50) for the family of PAH, namely polychlorinated dibenzofurans (PCDF), poly-halogenated dibenzo-p-dioxins (PHDD) and polychlorinated biphenyls (PCB) are taken as dependent variables and the HF energy (E), along with DFT-based global and local descriptors, viz., electrophilicity index (Ω) and local electrophilic power (Ω+) respectively are taken as independent variables. Fairly good correlation is obtained showing the significance of the selected descriptors in the QSAR on toxins that act as electron acceptors in the presence of biomolecules. Effects of population analysis schemes in the calculation of Fukui functions as well as that of solvation are probed. Similarly, some electron-donor aliphatic amines are studied in the present work. We see that global and local electrophilicities along with the HF energy are adequate in explaining the toxicity of several substances, both electron donors or acceptors when they interact with biosystems, in gas as well as solution phases.