I'm actually working on something like this now. It is likely that you won't get it all in one extraction anyway so it doesn't matter so much that you can't remove all of the residual first extract. Just do the best you can. You will probably end up pooling your extracts in the end.
I am using a water soluble solvent (methanol). I extract once (mix sample + 10 mL MeOH in a centrifuge tube, close the tube and sonicate for 10 minutes), centrifuge the sample to get the solid material to go to the bottom of the tube, thus making it easier to remove the solvent with a glass Pasteur pipette. Don't disturb the bed of plant matter as you remove the solvent. Repeat and keep the individual solvent layers separate. My plant material is very concentrated in "oil" so I take 1 mL aliquots of the individual extracts and dilute them separately to 100 mL with water. When I analyze for the volatiles in the solvent/water mixes, I see successively smaller amounts of the analytes in each extract. I carry this out until I get what I want out (usually 99% on the most difficult material to extract). That tells me how many times I need to extract the sample to get quantitative recovery. I was able to get 99%+ in 3 extractions (0.26 g plant matter and 10 mL shots of methanol). I pool the 3 for subsequent, more comprehensive analyses.
If you have a particular target molecule that is more easily extracted and you don't really care about the rest, then you might not have to extract as many times. Some analytes in my mix are no longer detectable after the first extraction but I can still detect them in the pooled extract. Good luck!