Soft rot is a commonly occurring potato tuber disease that each year causes substantial losses to the food industry. Here, we explore the possibility of early detection of the disease via gas/vapor analysis, in a laboratory environment, using a recent technology known as FAIMS (Field Asymmetric Ion Mobility Spectrometry). In this work, tubers were inoculated with a bacterium causing the infection, Pectobacterium carotovorum, and stored within set environmental conditions in order to manage disease progression. They were compared with controls stored in the same conditions. Three different inoculation time courses were employed in order to obtain diseased potatoes showing clear signs of advanced infection (for standard detection) and diseased potatoes with no apparent evidence of infection (for early detection). A total of 156 samples were processed by PCA (Principal Component Analysis) and k-means clustering. Results show a clear discrimination between controls and diseased potatoes for all experiments with no difference among observations from standard and early detection. Further analysis was carried out by means of a statistical model based on LDA (Linear Discriminant Analysis) that showed a high classification accuracy of 92.1% on the test set, obtained via a LOOCV (leave-one out cross-validation).