You can run a std alkane in GC-MS with the same prog as your sample run. Check the library too

Yes run an alkane mixture to help you calculate the retention indices of your unknowns. With this information you can filter out the MS library matches that don't fit both criteria, I.e high MS similarity match and close retention index values . However If your GC column is something other than the standard non polar HP-1ms then you may need to refer to the literature to find the retention indices for compounds you are identifying with the MS library you have. If you do not have access to relevant papers then try the website www.Chemspider.com as this lists retention indices for many compounds that have been separated on a variety of different GC columns.

From that spectra, you may take evaluation parameter as molecular wt and retention time. (refer The spectral interpretation book)

Dear Mahesh have you checked this peaks into standerd library, if it is not matching in standerd library then you need to consider other analytical techniques also becouse gc-ms only analyses the organic compound which are volatile.

Mr, Sohel, thanks for your comments, but i have already check library and there is no doubt GC-MS for volatile compounds. But my sample RT and mol. wt matches with more than one compound in library. That's why i asked this question.

Any one can explain me how can i predict compound which similar with more than one compound in the library, based on their RT and Mol. wt.

So you can consider the most close match

Dear Dr Patil

Just identify compound on the basis of mass fragmentation pattern

Dear

It's really need to consider many factors especially when the suggested compounds by library have close similarity. In this respect I suggest to consider the type of your samples and previous research done on it. However check the possible fragments could occur in your compound you may exclude one of suggested compounds. 

I hope you all pest, good luck

Thanks to every one for valuable suggestions.