This research serves as a feasibility study in assessing the applicability of machine learning techniques to detect fungal infections in apple crops. Fungal infections are a key challenge for the harvested crops. With marketing periods lasting up to several months for crops stored under controlled atmosphere in a storage facility, the undiagnosed disease development can have a significant impact on fruit security and harvest profitability, as well as result in food waste. The commonly used best practice employs laboratory-based polymerase chain reaction (PCR) testing, which can be expensive and time-consuming. This research employs mass spectrometry as a rapid alternative to PCR testing and examines the effectiveness of machine learning techniques to identify pathogen-specific biomarkers and to detect the presence of infections in apple crops. The analysis of variations of prevalent infections is provided, and the classification performance of various machine learning models is discussed. Support Vector Machine and AdaBoost provided 90 + % accuracy in multiclass classification of fungal infections. This research also identified that the species-specific key biomarker profile can be generated by incorporating between eight to twelve principal components. The rapid diagnostic and biomarker identification could be a benchmark for the control of fungal diseases in apple crops prior to entering storage facilities.