In DFT optimisation output, the dipole moment vales, Mulliken charges are displayed in two places, these are slightly differs from each other. Which one is correct? Any of related experts please suggest me sir
You didn't specify the software package, so its impossible to tell in an accurate manner. Usually some electronic properties are calculated for the input geometry (first set of data) and after the opt cycles they are calculated for the final geometry (second set of data).
Building up on Lucas' answer (and assuming the Gaussian software was used), population analyses are calculated for each step of the optimization process. If you have an opt job that required 20 steps to reach convergence, there will be 20 sets of atomic charges and dipoles (as well as many other properties), one for each of these steps. Consider using the last one, which will be the set for the optimized structure (assuming it reached convergence).