Hi Zhenbin,
What are your standards? Why can't you create a single contrived sequence that will produce them all from in silico digestion? Are they non-tryptic and you otherwise want to use tryptic digestion? Are you saying that your PD and MaxQuant searches are not currently finding all of your iRT standards? You will really need matched MS/MS spectra to your iRT standards in every run, if you want the most automatic support in Skyline to work for you.
Alternatively, if you can build up a fairly comprehensive iRT library before you try to import your searched fractions, if Skyline fails to find enough information about the standards in your library but is able to find enough run-to-run overlap (possibly challenging with fractions) and library-to-run overlap, it will use a regression between matched peptides and (non-standard) library peptides to derive iRT values, which just need to be derived relative to other peptides in our library.
If you can provide more information or example files, I may be able to help more.
--Brendan