Abstract: [Objective]The objective of this study is to discuss the feasibility of using artificial neural network for maize above-ground dry biomass estimation based on unmanned aerial vehicle (UAV) hyperspectral remote sensing imagery. [Method]This study took corn as study material, nitrogen fertilizer experiment was conducted in Caijia Town, Jilin Province. In the experiment, five nitrogen treatments and three replications were applied. UAV hyperspectral remote sensing data and above-ground biomass data were acquired for 2 critical growth stages of maize：V5-V6 and V11 (Ritchie growth stage). At last, 30 groups of data were collected, the maize above-ground dry biomass estimation model were designed using the acquired data. 22 groups of data were randomly selected from the all 30 groups of data to build the inversion model, and the remainder 8 groups of data were used to validate the performance of the model. The biomass estimation models were constructed based on spectral index method and BP neural network method respectively, and then they were compared. [Result]The results showed that the BP neural network model performed better than models designed by spectral indices. BP neural network model had a R2 value of 0.99, a RMSE value of 0.08 t/ha and a RMSE% value of 3.39% during calibration and had a R2 value of 0.99, a RMSE value of 0.15 t/ha and a RMSE% value of 8.56% during external validation. [Conclusion]The BP neural network model can effectively improve the accuracy of UAV’s hyperspectral remote sensing for inversion of maize biomass.