When we interact any 2D sheet with small molecule (eg. H2O, O2 ), Transnational vectors of whole system changes after optimisation. So can we consider for post calculation like PDOS plot, charge density visualisation , NEB ???????
Assuming you mean "translational vectors", i.e. the lattice vectors, then yes you should usually optimise them and use the optimised lattice for all your visualisation and analysis. I expect this is the correct approach for your molecule adsorbed on a 2D sheet.
There are some exceptions to the above rule, however. One common case is if you're studying a surface of a (3D) bulk material, where it is usually best to use the lattice vectors which were optimised for the perfect 3D bulk, and not optimise your actual surface calculation. This is because the material near the surface will relax differently and different lattice vectors will optimise its structure, it might even reconstruct into a completely different local structure. If your simulation had lots of bulk material under the surface then this wouldn't be an issue, but you can rarely afford to simulate so much material so the surface is a disproportionately large amount of the simulation and will affect the "optimal" lattice vectors to an unphysical degree.