By using different target prediction softwares I am getting lot of targets for a single miRNA which further vary on the basis of software used. So how to finalize a specific target to further validate it experimentally ?
It is not easy to get the precise target , but what you can do is take the common ones you found using all the software, and then try to validate them via WB/ other technique. Once validated by them you can go for a long term experiment
I assume that you have a hypothesis that you are trying to test so you may have some idea about the process that the miRNA is involved in.
Actually, there is nothing biologically wrong with a miRNA regulating 100s of targets (to our current knowledge, anyway) so the question is rather which of the targets are relevant to your hypothesis, I guess.
In this case you could use BLAST2Go to get more information about the common targets (see Debasish Paul) and then concentrate on the targets that fit to the question you are asking.
Keep in mind, that many of the targets are only relevant at a specific developmental stage, a specific tissue or in response to some stimulus.