Sensor-based detection of pests and pathogens in a high throughput and non-destructive manner is essential for mitigating crop loss. Infrared (IR) sensors in the form of vibrational spectroscopy provide both biochemical information about disease, as well as a large number of variables for chemometrics. This approach is highly adaptable to most biological systems including interactions between plants and their environments. Fast-acting necrotrophic fungal pathogens present a specific group of pests with adverse effects on food production and supply and are therefore pertinent to food security. Botrytis cinerea and Solanum lycopersicum are models for the study of fungal and crop biology respectively. Herein we use a compact mid-IR spectrometer with attenuated total reflection (ATR) attachment to measure the plant-microbe interaction between S. lycopersicum and B. cinerea on leaves, in vivo of intact plants. Chemometric models including exploratory principal component analysis (PCA) solely, and as a classifier in combination with linear discriminant analysis (PCA-LDA) are applied. Fingerprint spectra (1800−900 cm−1) were excellent discriminators of plant disease in both visually symptomatic as well pre-symptomatic plants. Major biochemical alterations in leaf tissue as a result of infection are discussed. Diagnostic potential for automatic decision-making platforms is shown by high accuracy rates of 100 % for detecting plant disease at various stages of progression.