Present studies have shown that near-infrared spectroscopy (NIRS) could be used to identify the producing area of apple. However, we still do not know whether the producing area affects the optical properties of apple, whether it affects the relationship between optical properties and internal qualities, and whether the optical parameters could be used to predict the apple-producing area. To answer these questions, the apples produced in three areas with different temperatures, precipitations, and altitudes in the Loess Plateau region, China, were used as samples to obtain the optical properties of apple pulp by a single integrating sphere system, analyze the differences in optical properties of apples produced in different areas, analyze the correlations between optical properties (absorption coefficient (μ a) and reduced scattering coefficient (μ s ')) and internal quality indices (soluble solids content (SSC), moisture content and firmness), and investigate the feasibility of discriminating the apples produced in different areas and predicting the internal qualities using NIRS. The results showed that the apples produced in Luochuan had the highest SSC while Yangling apples had the highest moisture content. The μ a had three absorption peaks at around 985, 1200, and 1430 nm, and the μ s ' of apples from three areas showed a similar trend with wavelength. The best correlation of μ a with SSC and moisture content was found in Yangling apples with a correlation coefficient of −0.90 and 0.79, respectively. The built linear discriminant analysis based on μ a spectra performed better in discriminating producing area. The established partial least squares regression models based on μ a spectra had the best prediction performance for SSC and moisture content, while the model built with μ s ' had the best prediction performance for firmness. This study is helpful to understand the differences in optical property of 'Fuji' apple produced in different areas and offers information for identifying the producing area of fruit using NIRS technology. • The internal qualities of apple produced in three areas had significant difference. • The apples produced in three areas had three absorption peaks at 980, 1200, and 1430 nm. • The best correlation of μ a with SSC and moisture content was found in Yangling apples. • μ a spectra performed better in discriminating the apple-producing area. • μ a and μ s ' spectra had the best prediction performance for SSC and firmness, respectively. [ABSTRACT FROM AUTHOR]