What do you mean? Methods? Program packages? If you are searching for the best method, you need to look at first at the system size. Is it possible to run calculations on high level? Do you want to optimize the geometry of the excited state? Do you need electron difference densities between the ground and excited state to see how the electrons are reordered in the excited state? You really should make it more precise if you want to get help!
I am not sure if semi-empirical methods are that good for transition metal complexes. Here the parametrization of the metals plays a very important role. I would start from first principles...
MINDO/d is quite suitable for transition metal complexes. But you could use DFT (Density Functional Theory) if you could select the appropriate basis set for the selected transition metal.
MINDO/d is availabe with other semi-imperical methods like AM1, MINDO/3 and so on in the same package
you could see the following articles
A. V. Soudackov, A. L. Tchougreeff and I. A. Misurkin,
Theoretical Chemistry Accounts: Theory, Computation, and Modeling (Theoretica Chimica Acta)
Volume 83, Numbers 5-6, 389-416
W. Th. A. M. Van Der Lugt, Molecular-orbital calculations on transition metal complexes, charge-transfer spectra and the sequence of metal and ligand orbitals, International Journal of Quantum Chemistry
Very nice! Thanks for that. Unfortunatly there seems to be no large benchmark available. But according to the results of Soudackov et al. this method seems to be very suitable for excitation spectra transition metal complexes. But I'm still not convinced that this method does not have the same problems with groundstate geometries as INDO/S. If that is the case for MINDO/d, than it would not be suitable for the calculations Badr Elsayed wants to to carry out.