This study delves into the detection of the mycotoxin zearalenone (ZEA) in popcorn, aligning with the broader goal of ensuring food safety and security. Employing fast, non-destructive near-infrared spectroscopy, the research analyzes 88 samples collected in France. In order to emphasize the dedication to robust methodologies, an essential element of sustainable practices, the assessment of various validation methods becomes significant. Six CART classification tree models, with a threshold of 68 µg/kg, are meticulously assessed. The study not only scrutinizes various validation strategies but also explores the concrete impact of the detection process, emphasizing sustainable practices. Model F (Kennard and Stone) is chosen for its commendable ability to generalize and its balanced performance, boasting 91% precision and 57% recall. Notably, this model excels in specificity, minimizing false positives and contributing to food safety. The identification of key wavelengths, such as 1007 nm, 1025 nm, and 1031 nm, highlights the potential for targeted interventions in crop management. In conclusion, this research showcases near-infrared spectroscopy as a sustainable approach to fortifying the food safety of popcorn, paving the way for advancements in ZEA risk detection and prevention, while minimizing environmental impact.