Although BTP (2,6-di(1,2,4-triazin-3-yl)pyridine) has been widely evidenced as the most effective nitrogen ligand for the selective complexation of trivalent actinides from lanthanide counterparts, the origin of its selectivity is still an open question. Neither experimental data nor theoretical calculations have been able to rationalize the role of covalency in real experimental BTP complexes. We show herein with DFT calculations on [M(BTP)(3)](3+) (M = La, U, Cm, Gd) that, even if back-bonding effects are significant in the U-BTP bond, it is the contrast of donation on 6d and 5f Cm-III orbitals that explains, at least in part, its selective complexation to BTP.