Genetic algorithm interval partial least square (GA-iPLS) and genetic algorithm partial least square (GA-PLS) were proposed to select the characteristic wavelength region and characteristic wavelength of sugar content against apple near-infrared spectra for sugar content prediction. The apple near-infrared spectra data were divided into 40 intervals. Consequently, 5 subsets (No.4,6,8,11,18) and 362 data points were selected quickly by GA-iPLS, and 44 characteristic wavelengths were selected by GA-PLS based on the 5 subsets. Compared with the whole spectra data model, the GA-iPLS and GA-PLS models could not only improve precision with the coefficients of determination for prediction set improved by 10%, but also simplify the model with 7 primary factors decreased in the model. With the proposed methods, a concise easily computed model can be built to select the characteristic reigon and wavelength of near-infrared spectra.