We report results of Broken Symmetry Density Functional Theory (BS-DFT) calculations providing the optimized geometries and the exchange coupling constants for three members of the quasi-onedimensional manganese-porphyrin family of molecular magnets, [MnTPP][TCNE], [MnOEP][HCBD] and [MnTtBuPP][HCBD] (TPP = meso-tetraphenylporphyrinato, OEP = octaethylporphyrinato, TtBuPP =meso-tetrakis-(4'-tert-butylphenyl)porphinato, TCNE = tetracyanoethylene, and HCBD = hexa-cyanobutadiene). We compare the results of BS-DFT calculations for extended systems, with periodic boundary conditions, and for finite systems, magnetic dimers modeling the actual molecular magnets. By varying systematically the main angles, we are able to determine the geometry dependence of the exchange interaction. Structureproperties correlations in these charge transfer salts reveal the determinant role of the Mn- (N≡C)TCNE bond angle on the strength of the ferrimagnetic coupling between the MnIII and the TCNE spins. The large differences between the magnetic properties of these systems are explained based on the correlation between the exchange coupling constant and the overlap between the metal ion and ligand orbitals [1,2]. We also report quantum chemical calculations to explain the ferrimagnetic ordering as well as the structural and magnetic disorder of V[TCNE]x (x~2), the first room-temperature molecule-based magnet. Starting from an ideal lattice model containing TCNE ligands either tetra- or bi-connected to V(II) ions, we optimize the geometry for the ideal lattice using DFT calculations with periodic boundary conditions. Broken symmetry DFT calculations indicate antiparallel spin alignment resulting in ferrimagnetic ordering, but heavily overestimate the value of the exchange coupling. Better estimates of the exchange coupling parameters between the V(II) ion and the [TCNE].- anionic radical are obtained by means of multiconfigurational calculations performed on smaller molecular models cut from the optimized crystal lattice. We also identify the key sources of structural disorder, explaining the amorphousness and non-stoichiometric nature of V[TCNE]x. The broken symmetry DFT approach evidences ferrimagnetic spin orientation for all TCNE rotated configurations, ruling out the spin glass model. CASSCF calculations with additional spin orbit interaction allow for the account of the single-ion-anisotropy of the V(II) ions in different environments. We determine a small uniform zero-field-splitting of the bulk as well as a sizeable random anisotropy due to TCNE vacancies [3,4].