I was trying to dock a ligand to 2 viral protein at first and got good binding properties at very important binding sites of virus, but when I tried to dock with other 15 viruses, saw type of exceptional binding was found. Can this be possible?
Yes this is possible. Perform multiple sequence alignment for 17 viral proteins (15+2) and inspect the residue columns of the MSA corresponding to the amino acids present in the binding site. Check the viral proteins genus/family to understand how the ligand binding site may be evolutionarily modified. These are sequnce-based approaches. You may also compare the binding sites of 17 viral proteins using 3D/structure-based approaches. Please make sure that you always use ligand bound viral protein structures taken from the PDB. These tasks will provide you some idea why there is a subtle change in the properties of ligand-binding cavities.
Multiple sequence alignment will definitely gives you an idea about this exceptional similarity, another thought about structure approaches is to superimpose these structures if you have or you can generate their homology models and calculate RMSD at binding residues of your ligand.