Abstract: The quality of Aztec apples varies significantly at different maturity stages, which can have a significant impact on postharvest storage and sales efficiency. This study focused on Aztec apples at four different maturity stages in Suqian, Jiangsu Province. Firstly, the variations in color (L^*, a^*, b^* values), firmness (FI), soluble solid content (SSC), titratable acidity (TA), moisture content (MC) and dry matter content (DMC) were analyzed using principal component analysis (PCA) and linear discriminant analysis (LDA). Simultaneously, visible and near-infrared (Vis-NIR) and near-infrared (NIR) spectral techniques, along with the successive projections algorithm (SPA), competitive adaptive reweighted sampling (CARS) and uninformative variable elimination (UVE) algorithms were employed for selecting relevant characteristic variables. Subsequently, partial least squares (PLS) and support vector machine (SVM) were utilized to establish quality prediction models for Aztec apples. The results revealed that SSC, a^*, L^* and b^* had a significant impact on the categorization of Aztec apples at different maturity stages. Notably, wavelength bands in the ranges of 510 to 680 nm, 1170 to 1270 nm and 2300 nm exhibited strong correlations with characteristic attributes. The SPA-PLS and SPA-SVM models demonstrated remarkable performance in predicting the L^*, b^* and a^* values of Aztec apples at different maturity stages, with all relative percent deviation (RPD) values exceeding 3.00. The CARS-PLS model effectively predicted SSC with an RPD of 3.19. However, the prediction accuracy of SPA-PLS models for FI, TA, MC and DMC was comparatively lower, with RPD values of 2.27, 2.21, 2.32 and 2.42, respectively. The results demonstrated that Vis-NIR and NIR spectroscopy methods could predict the quality of Aztec apples at different maturity stages, providing valuable technical references for the harvest management and quality control of Aztec apples.