The spread of the typical poisonous grass Aconitum leucostomum destroys the spatial distribution structure of grassland populations, degrades grasslands and affects the development of animal husbandry. Remote sensing monitoring and spatial and temporal distribution characteristics analysis of Aconitum leucostomum in the Narathi grassland of Xinjiang are important for the protection of grassland ecological security and sustainable development of animal husbandry. In this study, the 2021 UAV images of Narathi grassland and the 2013-2022GF-1 satellite images were used as data sources, and the random forest (RF), neural network (NN) and decision tree (DT) methods were used to identify the Aconitum leucostomum in the 2021 UAV remote sensing images. The three models, Dominant class variability-weighted method (DCVW), Local average method (LA) and Cubic Convolution Interpolation method (CCI), were used to classify and identify the Aconitum leucostomum in 2021 UAV remote sensing images. The spatial upscaling of the UAV image (0.01m) to the same scale as that of the GF-1 satellite image (16m) was performed by fitting the reflectance of the two images at the same position to construct a calibration model of the GF-1 satellite image from 2013 to 2022 and analyze the spatial and temporal dynamics of the Aconitum leucostomum. The results show that (1) the decision tree method has the best classification with an overall accuracy of 90.88% and a Kappa of 0.87. (2) The Aconitum leucostomum identified by the UAV image was spatially up-scaled, where the CCI method has the best conversion. The reflectance of the same position of the UAV image and GF-1 satellite image was fitted with Gaussian function, and the $\mathrm{R}^{2}$ was 0.2851 and the RMSE was 0.0011. (3) The growth area of Aconitum leucostomum within the Narathi grassland generally showed a fluctuating growth trend from 2013 to 2022, with an increase of 3170.08 ha mainly in the eastern part of the Narathi grassland. The integrated application of UAV and GF-1 satellite remote sensing realized high-resolution and large-area monitoring of Aconitum leucostomum, which provided reference information for the integrated control of Aconitum leucostomum.