A number of hexacoordinate, pentacoordinate, and tetracoordinate Ni(II) complexes have been investigated by applying ab initio CASSCF + NEVPT2 + SOC calculations and Generalized Crystal Field Theory. The geometry of the coordination polyhedron covers D4h, D3h, D2h, D2d, C4v, C3v, and C2v symmetry. The calculated spin-Hamiltonian parameters D and E were compared to the available experimental data. The limiting values of the D-parameter in the class of Ni(II) complexes are identified. Magnetic anisotropy in Ni(II) complexes, expressed by the axial zero-field splitting parameter D, seriously depends upon the ground and first excited electronic states. In hexacoordinate complexes, the ground electronic term is nondegenerate 3B1g for the D4h symmetry; D is slightly positive or negative. In tetracoordinate systems, D is only positive when the electronic ground state is nondegenerate 3A or 3B; this diverges on the τ4 path when oblate bisphenoid approaches the prolate geometry and a level crossing with 3E occurs. In pentacoordinate systems, D could be extremely negative when approaching a trigonal bipyramid (Addison index τ5 ∼ 1, ground state 3E″). In pentacoordinate Ni(II) complexes with the D3h and C3v symmetry of the coordination polyhedron, the ground electronic term is orbitally doubly degenerate which causes the D-parameter stays undefined. It is emphasized that one has to inspect compositions of the spin-orbit multiplets from the spin states |MS⟩ and check whether the weights confirm the expected spin-Hamiltonian picture: with D > 0, the ground state contains a dominant part of |0⟩ (close to 100%) whereas with D < 0 the spin-orbit doublet is formed of |±1⟩ with high weights (approaching 50 + 50%). The calculations show that the situations are not black and white, and the mixing of the states might be more complex especially when the rhombic zero-field splitting parameter E is in the play. In the case of the 3E ground term, six spin-orbit multiplets are formed by mixing six |MS⟩ states from the ground and quasi-degenerate excited states.