One hundred and twenty citrus trees under three cultivation patterns including wide row and narrow plant，wide row and narrow plant fence pattern and traditional cultivation were used to measure the content of nitrogen in the canopy and extract the texture index and vegetation index from the multispectral images data of UAV remote sensing to quickly and accurately obtain the content of nitrogen and spatial distribution characteristics of plant canopy，and to manage the large-scale orchard accurately and dynamically.The random forest （RF） algorithm was used to establish the inversion model of nitrogen in the citrus canopy based on vegetation index，texture index，and the integration of vegetation index and texture index.The inversion accuracy of support vector machine （SVM），BP neural network algorithm （BP），and RF models that integrate vegetation index and texture index was compared.The results showed that the integration of vegetation index and texture index predicted the content of nitrogen in citrus canopy more accurately than the single vegetation index or texture index in the random forest algorithm.The training sets R² and the test sets R² of the vegetation index，texture index，and integration of vegetation index and texture index were0.710 and 0.430，0.761 and 0.349，0.775 and 0.533，respectively.The training sets R² and the test sets R² of the integration of vegetation index and texture index in the SVM algorithm and BP neural network were0.511 and 0.371，0.651 and 0.204，respectively.The results of using the RF model of vegetation index and texture index to inverse the content of nitrogen in citrus orchards under three cultivation patterns showed that the average content of nitrogen in citrus canopy in wide row and narrow plant werethe highest，followed by the wide row and narrow plant fence pattern，and the traditional cultivation pattern was the lowest，with the average content of nitrogen being 31.33，30.20，and 27.82 mg/g，respectively.It is indicated that the random forest algorithm combining UAV remote sensing with vegetation index and texture index can effectively predict the content of nitrogen in citrus canopy.It will provide a reference for the quantitative fertilization of large-scale citrus orchards.