Abstract: Nitrogen is one of a large number of elements in process of plant growth and development．The rapid and accurate acquisition of nitrogen content is of great significance for crop monitoring and management in the field．Leaf Nitrogen Content（LNC）of buckwheat canopy was quantified by UAV equipped with multispectral sensors．It could provide a theoretical basis for information management of buckwheat leaves．Taking different buckwheat varieties as research object, by using UAV to obtain multispectral images of buckwheat canopy leaves during flowering and filling stages, and synchronously collect LNC of buckwheat canopy leaves, reflectivity was extracted at 5 bands．12 vegetation indices related to LNC were selected for Pearson correlation analysis．Characteristic variables with higher correlation among 17 spectral variables were selected for PLSR, SVM, and BPNN regression modeling with measured LNC．G, R, NIR, NDVI, RDVI, RVI, SAVI, NLI, OSAVI, GRVI have a high correlation with LNC, with the highest being GRVI, reaching 0.824．Regression model established by BP neural network showed the best performance．Coefficient of determination（R²）and root mean square error（RMSE）of prediction set were 0.828 and 2.172, respectively．The R², RMSE and RPD of validation set were 0.939, 1.100 and 4.587 respectively．Therefore, the UAV multi-spectral remote sensing technology could realize field-scale LNC estimation of buckwheat canopy leaves．