Lignocellulosic biomasses have some characteristics that hinder their use as fuel. Further, the biodegradation processes that can occur during storage can affect their energy properties. Torrefaction can improve their energy properties while reducing the impacts of storage. Near-infrared (NIR) spectroscopy could be a promising tool to estimate changes to biomass properties during storage. A series of waste feedstocks, including: sugarcane bagasse, coffee husk, eucalyptus, and pine were torrefied at 290 °C in a screw reactor, over 5, 7.5, 10, 15, or 20 min. Then, raw and torrefied biomasses were submitted to leaching and to white and brown-rot fungi, to simulate storage conditions. The weight loss (WL) due to fungal degradation after 2, 4, 8, and 12 weeks was measured and the high heating value (HHV) of all biodegradation steps, including raw and torrefied samples after leaching, were determined. All the samples were analysed by NIR spectroscopy and chemometric models were then developed. Firstly, partial least squares for discriminant analysis (PLS-DA) models successfully classified the four different residual biomasses into raw and torrefied forms according to fungal decomposition. Secondly, partial least squares regression (PLSR) models showed potential utility in an industrial context as a standardized continuous method to predict the HHV during biomass storage steps. While, PLSR models did not present good accuracy when estimating WL resulting from fungal degradation, they can be useful for screening during decision making. Further studies are required to improve and develop more efficient models to predict the fungal degradation level of stored biomasses. For exemple, in considering the class of torrefied biomasses, the model could show better predictive capacity due to less variability in the data. These results highlight the potential of NIR spectroscopy as a simple, fast, and efficient tool to analyze the degradation process over time. Such a rapid and non-d