Then is not the same compound if it has different retention time. When doing GC-MS there are two things that characterizes a compound, the ion fragments and the retention time.
When dealing with complex matrices it can occur that you find interferent peaks that has the same ions as your compound but different retention times. To know how to select the correct peak, the only way is to have a reference standard of the compound of interest.
Then the SI and RSI can be indicative of how similar is the spectra that you have compared to the NIST library. You can also use this as key to know which one of your peaks matches best.
You do not write what components are involved. However, there is sometimes the situation that positional isomers cannot be distinguished in a library search. As an example I mention the xylene isomers, on my column m- and p-xylene elute at the same time, o-xylene a little later. In the library search all 3 isomers have a match of 95-99 %, independent of the isomer. That means also at the retention time of o-xylene, m-,p- and o-xylene are shown with a match of 95% or more. Unfortunately, the library cannot distinguish between these isomers. It is possible that a similar phenomenon exists in your case.
It might be helpful to first name the coponents and possibly show a mass spectrum and chromatogram. This would make it much easier to identify the causes.
Here you could be dealing with a number of different issues. First, are you using the same column? If not, you'll get high retention time with longer column and vice versa with shorter column. Second, is your pressure the same? If your pressure is varying, the samples will give different RT. Third, is your column temperature the same? If all these are checked and you're sure no discrepancies exist then I'll probably say check your sample prep.
Would you please supply more details? Did you derivatize your samples? Derivatization by methoximation/silylation always generates two geometric isomers of monosaccharides.