Interactions of Fusarium species with different wheat varieties differ in their temporal dynamics and symptom appearance. Reliable and objective approaches for monitoring processes during infection are demanded for plant phenotyping and disease rating. This study presents an automated method to phenotype wheat varieties to Fusarium head blight (FHB) using hyperspectral sensors. In time-series experiments, the o p t i c a l p r o p e r t i e s o f s p i k e s i n f e c t e d w i t h F. graminearum or F. culmorum were recorded. Two hyperspectral cameras, in visible and near-infrared (VIS-NIR, 400–1000 nm) and shortwave-infrared (SWIR, 1000–2500 nm) captured the most relevant bands for pigments, cell structure, water and further compounds. Correlations between disease severity (DS), spike weight, spectral bands and vegetation indices were investigated. Following, the detectability of infections was assessed by Support Vector Machine (SVM) classifier. A variety ranking based on AUDPC was performed and compared to a fully-automated approach using Non-metric Multi-Dimensional Scaling (NMDS). High correlation was found between the spectral signature and DS in 430–525 nm, 560–710 nm and 1115–2500 nm. All indices from the VIS-NIR showed high correlation with DS and, for the first time, this was also confirmed for three indices from the SWIR: NDNI, CAI and MSI. Using SVM, differentiation between healthy and infected spikes was possible (acc. > 0.76). Furthermore, the possibility to differentiate between F. graminearum and F. culmorum infected spikes has been verified. The NMDS approach was able to reproduce accurately the variety ranking and outlines the potential of hyperspectral imaging to phenotype the variety susceptibility for improved breeding processes.