Abstract: Abstract: Hyperspectral remote sensing has widely been used to estimate crop physiological, ecological, and biochemical parameters in recent years. However, most previous studies focused mainly on the selection of sensitive bands or the construction of vegetation index (the combination of sensitive bands) for crop parameter inversion. Particularly, the spectral information of some bands can be lost, and then to reduce the prediction ability of the estimation model. The purpose of this study is to estimate the aboveground biomass and leaf area index of summer maize using all spectral information (spectral bands). Therefore, a three-year (2018-2020) field experiment was also conducted under different water and nitrogen management in the Guanzhong Plain of China. Accordingly, 212 plant samples (aboveground biomass and leaf area index) were collected during the vegetative growth period of summer maize. Prior to plant sample collection, the hyperspectral reflectance data of the summer maize canopy was measured using an ASD FieldSpec 3 portable spectroradiometer. Correspondingly, the estimation model was constructed using Partial Least Squares Regression (PLS), Extreme Learning Machine (ELM), Random Forest (RF), and Stacked Ensemble Extreme Learning Machine (SEPLS_ELM, using the PLS stacked ensemble strategy). The results showed that the estimation accuracy (four estimation models) of the leaf area index of summer maize was higher than that of aboveground biomass. The estimation models of PLS and ELM presented a relatively low accuracy for the aboveground biomass and leaf area index of summer maize, where the determination coefficient (R2) for the validation set of the aboveground biomass estimation model was lower than 0.85, and the Root Mean Square Error (RMSE) was higher than 550 kg/hm-2, whereas, the R2 for the validation set of leaf area index estimation model was lower than 0.90, and the RMSE was higher than 0.40 cm2/cm2. The estimation model of aboveground biomass and leaf area index of summer maize using RF and SEPLS_ELM presented a higher estimation accuracy, particularly that the performance of the SEPLS_ELM model was outstanding. The R2 values for the validation set of aboveground biomass and leaf area index estimation model using the SEPLS_ELM model were 0.955 and 0.969, while the RMSE were 307.3 kg/hm2 and 0.24 cm2/cm2, and the Residual Predictive Deviation (RPD) were 4.66 and 5.30, respectively. Compared with PLS and ELM, the estimation accuracy of the SEPLS_ELM model was significantly improved (the R2 increased by more than 8%, RMSE decreased by more than 40%, and RPD increased by more than 70%, respectively) in aboveground biomass and leaf area index estimation. Compared with the RF, the R2 of the SEPLS_ELM estimation model increased by more than 7%, RMSE decreased by more than 40%, and RPD increased by more than 66% in the aboveground biomass and leaf area index estimation of summer maize, respectively. Consequently, the present study demonstrated that the SEPLS_ELM model was highly reliable to predict the aboveground biomass and leaf area index of summer maize. The findings can provide a strong reference for the estimation of crop aboveground biomass and leaf area index using hyperspectral remote sensing.