The relevant growth of the wood pellet market in Europe in the last decade led to an increased focus on solid biofuel as a necessary and available renewable resource for energy production. Among biofuels, wooden pellets are among the most widespread for domestic heating. Therefore, monitoring the qualitative properties of commercialized pellets is crucial in order to minimize the amount of harmful emissions in residential areas. Standard ISO 17225 sets threshold values for the chemical and physical properties that commercialized biofuels must fulfil. Specifically, ISO 17225-2 defines that pellets for residential use must be produced from virgin wood, but no method is proposed to assess the actual origin of the material, leading to the risk of the commercialization of pellets made up from chemically treated materials. This study proposes a model obtained via near infrared spectroscopy analyses and chemometrics methods, such as classification, to rapidly assess whether pellets are made up of virgin or chemically treated wood. The result suggests the effectiveness of NIRs for the detection of non-virgin pellets with an accuracy greater than 99%. Furthermore, the model appeared to be accurate in the assessment of both milled and intact pellets, making it a potential in-line instrument for assessments of pellets’ quality.