The occurrence of stuck and sluggish wine fermentations is a persisting problem in the wine industry worldwide. This study illustrates the suitability of headspace solid-phase dynamic extraction coupled with gas chromatography-mass spectrometry (HS-SPDE GC-MS) for wine analysis and the subsequent application to discriminate between control and problem fermentations using partial least-squares discriminant analysis (PLS-DA) models. The specific analytical technique is relatively new and has not yet to the authors' knowledge been evaluated for the analysis of wine within this context of problem fermentations. HS-SPDE GC-MS was used to determine 68 volatile compounds (higher alcohols, fatty acids, esters, and carbonyl compounds) in 94 monovarietal fermenting must samples consisting of 56 red and 38 white cultivars. PLS-DA models showed the potential to discriminate between control and problem fermentations using corrected peak area headspace data for the 68 analytes. This possibility to discriminate between problem and control fermentations with only the headspace data could possibly be applied for the prediction of problem fermentations in future studies and to better understand the chemical causes of problem fermentations.