This research paper investigates the application of machine learning techniques for fruit ripeness detection using near-infrared (NIR) spectrum data. By leveraging a dataset comprising spectral information and key biochemical parameters of mangoes, including Vitamin C, total soluble solids, and total acidity, we developed and evaluated machine learning models to predict these parameters based on NIR spectral absorbance values. Our study employed Random Forest Regression, Support Vector Regression (SVR), and Partial Least Squares Regression (PLSRegression) models to analyze and predict soluble solids content—a critical indicator of fruit ripeness. The performance of these models was assessed based on root mean square (RMS) error metrics on both training and testing datasets. These results underscore the potential of machine learning algorithms, particularly Random Forest Regression, in non-destructive fruit quality assessment and ripeness detection tasks, offering valuable insights for optimizing agricultural practices and enhancing fruit quality management in the food industry.