Abstract: 【Objective】 Unmanned aerial vehicles (UAVs) equipped with digital cameras offer high mobility, flexibility, and resolution, making them advantageous for rapid and accurate AGB estimation. This study aims to investigate the impact of image resolution differences at various flight altitudes of UAVs on the accuracy of AGB estimation. 【Methods】 In this study, we conducted UAV flights at five different altitudes (10, 30, 50, 70, 90 m) over the Zhaosu mountain meadow in Xinjiang to explore the effects of image acquisition at different flight altitudes on AGB estimation accuracy by analyzing differences in spectral information and texture features. 【Results】 By extracting digital images collected at different flight altitudes, we analyzed their spectral information and texture features and correlated these features with measured AGB.The top 8 vegetation indices selected in this study and the top 8 texture features showed strong correlations with AGB. After integrating the three input variables’ variance inflation factor (VIF), we constructed an AGB estimation model using Principal Component Analysis (PCA) and Multiple Linear Regression (MLR). The study compared the model accuracy for estimating AGB using images of different resolutions. 【Conclusion】 The results indicate that: (1) The correlation between texture features and AGB is weaker than that of vegetation indices when image resolution ranges from 0.27 to 2.45 cm but both of them reach a significant level. As image resolution decreases, the difference in correlation between the two becomes more apparent. (2) At the same image resolution, the best AGB estimation results are achieved when spectral information is combined with texture features, followed by a model using a single texture feature, with a single spectral model performing the worst. (3) As image resolution increases, the accuracy of AGB estimation improves for models using spectral information, texture information, and spectral + texture information.