First reported at a nursery in Mpumalanga province in 1990, Fusarium circinatum is a fungal pathogen causing widespread mortality of Pinus radiata and Pinus patula seedlings. Improved methodologies for early disease detection are thus pertinent, and rely on identifying specific wavebands that correspond to specific physiological responses of the plant to stress. The objectives of this research were to i) determine the earliest possible window period, from time of infection, for disease detection, and ii) identify the specific hyperspectral wavebands that could be used for discriminating healthy and infected seedlings. To achieve these objectives, we setup an experiment with a sample of 3-month old P. radiata seedlings (n = 150) divided into three classes; healthy (n = 50), damaged (n = 50), and infected (n = 50). Reflectance measurements for all three classes were collected using an Analytical Spectral Devices Full Range spectroradiometer, weekly over a five week period. Reflectance measurements were later analysed using a random forest with a feature selection algorithm. Results of the analysis indicate that the best possible discrimination occurs at week three (KHAT = 0.81; out of bag (OOB) error = 12.67%). The results further indicate that wavelengths in the red-edge and near-infrared regions show the most promise in discriminating the healthy, damaged, and infected classes. These results could be explained by reduction in needle chlorophyll content expressed by a shift in the red-edge toward shorter wavelengths. Furthermore, lowered near-infrared reflectance has been associated with disease-induced stress. Overall, this study provides a basis for the early detection and discrimination of infected P. radiata seedlings that could be operationalized within a nursery environment.