This paper is devoted to the determination of the spin density in the S = 2 ground state of [Mn(Me(6)-[14]ane-N-4)Cu(oxpn)](CF3SO3)(2) with Me(6)-[14]ane-N-4 = (+/-)-5,7,7,12,14,14-hexamethyl-1,4,8,11-tetraazacyclotetradecane and oxpn = N,N’-bis(3-aminopropyl)oxamido. The crystal structure, previously determined at room temperature through X-ray diffraction and at 40 K through unpolarized neutron diffraction, consists of oxamido-bridged [Mn(Me(6)-[14]ane-N-4)Cu(oxpn)](2+) cations and non-coordinated triflate anions. Within the cation the S-Mn = 5/2 and S-Cu = 1/2 local ground states are antiferromagnetically coupled, which gives rise to a ground S = 2 and an excited S = 3 pair states, with a quintet-septet energy gap of -93.3 cm(-1). The experimental spin density has been deduced from polarized neutron diffraction data recorded at 2 K under 50 kOe. Positive spin populations were observed on the manganese side, and negative spin populations on the copper side. The delocalization of the spin density from the metal center toward the terminal and bridging ligands was found to be more pronounced on the copper side than on the manganese side, and the nodal surface (of zero spin density) is closer to the manganese than to the copper atom. The experimental data have been compared to the results of several theoretical approaches, corresponding to different levels of sophistication. These approaches are as follows : (i) the pure Heitler-London description of the S = 2 ground state; (ii) the incorporation of the spin delocalization in the Heitler-London scheme, using the concept of magnetic orbitals in the extended Huckel formalism; and (iii) two types of density functional theory methods. In cases ii and iii, the experimental spin populations have been fairly well reproduced. The DPT approaches have provided some important insights on both spin-delocalization and spin-polarization effects.