Rice (Oryza sativa L.) is unique among major cereal crops, as it is primarily consumed as a whole grain after cooking. Evaluating rice quality is crucial for maintaining high standards and meeting consumer expectations. Quality assessment involves multiple parameters, including appearance, texture, aroma, taste, nutritional content, and safety—factors that collectively influence the overall quality and market value of rice products. Near-infrared spectroscopy, combined with machine learning techniques, was employed to link molecular characteristics to quality traits, offering a high-throughput and efficient evaluation method. Partial Least Squares regression models demonstrated strong predictive performance for several key parameters, whiteness (R² = 0.94), grain width (R² = 0.94), resilience (R² = 0.96), and springiness (R² = 0.98), identifying important wavelength regions. Principal Component Analysis revealed clear clustering patterns among the rice varieties, while Partial Least Squares Discriminant Analysis achieved a 17% error rate in external validation. The accuracy for the training process was significant, registering a not-assigned value for the samples used for the calibration step (23%). The cross-validation process was characterized by an accuracy of 68%, an error rate of 21%, and a not-assigned value (28%). The cross-validation process was characterized by an accuracy of 68%, an error rate of 21%, and a not-assigned value (28%). Notably, spectral markers at A6032/4457 cm⁻¹, A7004/5241 cm⁻¹, and A7004/4749 cm⁻¹ reflected distinct biomolecular differences between varieties. These markers enable accurate quantification, classification, and differentiation of rice types, enhancing quality control, breeding selection, and consumer satisfaction. This study successfully developed a classification model using PLS-DA based on NIR spectroscopy data to distinguish rice varieties by their physicochemical properties. The high classification accuracy underscores the potential of integrating chemometric tools with NIR spectroscopy for advanced grain evaluation. Spectral markers capturing biomolecular traits prove to be powerful tools, improving sensitivity and efficiency, and reducing the time and resources required for comprehensive rice quality assessment.