Rapid, real-time, and accurate acquisition of the nitrogen nutrition status of winter wheat is crucial for evaluating winter wheat growth, estimating yield, and guiding agricultural modernization and production. This study utilized hyperspectral cameras and digital cameras mounted on drones to collect canopy spectral data during the three critical growth periods. Simultaneous ground experiments were conducted to determine the physical and chemical properties of biomass and plant nitrogen content. Four characteristic parameters, including the vegetation index, the red edge index, the red edge parameter, and the three-band parameter of the hyperspectral image, as well as the color index of the digital camera and its fusion parameters, were selected. Partial Least Squares Regression (PLSR), Stepwise Regression (SWR), Random Forest (RF), and Back Propagation (BP) algorithms were used to establish a winter wheat nitrogen nutrition index monitoring model. The accuracy of the model was evaluated, and the optimal estimation model was selected. The results showed that (1) In univariate modeling, the nitrogen nutrition index model constructed with the red-edge parameter DIDRmid was the best, achieving a modeling R² of 0.66, RMSE of 0.11%, and a validation R² of 0.55, RMSE of 0.13%. (2) In the multi variate modeling, the nitrogen nutrient index model constructed with the red edge parameter as the independent variable was superior to the nitrogen nutrient index model constructed with the vegetation index, the red edge index, the three band parameter and the color index as the independent variables, where the nitrogen nutrition index model constructed by the BP algorithm based on red edge parameters was optimal (modeling R²=0.75, RMSE=0.10%, validation R²=0.60, RMSE=0.12%).(3) In fusing hyperspectral parameters and digital index variable modelling, the nitrogen nutrient index model constructed with the multimodal variable red edge parameter + color index was superior to the nitrogen nutrient index models constructed with the red edge parameter + vegetation index, red edge parameter + red edge index and red edge parameter + triple band parameter, where the nitrogen nutrient index model constructed using PLSR with the fusion of the red edge parameters and color index was optimal (modeling R²=0.77, RMSE=0.09%, validation R²=0.65, RMSE=0.11%). Its model accuracy was better than that of univariate modeling and multivariate modeling. This study can provide an important reference for estimating the nitrogen nutrition status of winter wheat.