The Source region of Yellow River (hereafter referred to as the source region) is located in the transition zone between the permafrost and seasonal permafrost zones on the Qinghai-Tibet Plateau, and alpine shrubs, recognized as indicators of climate change, are widely distributed in the Source Region. Against the backdrop of climate change and permafrost degradation, the phenomenon of shrub expansion has attracted widespread attention. However, current studies are mostly confined to small scales, making it difficult to understand the temporal coverage status of shrubs across the entire region. This paper focuses on the cold alpine shrubs in the source region, using Landsat 8 remote sensing images and drone aerial photography data as the primary data sources.,and applying machine learning methods such as random forests, support vector machines, and artificial neural networks to retrieve the alpine shrub coverage in the source region from 2014 to 2024., mathematical statistics methods are used to analyze the spatial distribution patterns of shrubs combined with data of topography, meteorology, and other factors over the years in the source area. Methods such as the coefficient of variation (CV), Sen Mann-Kendall trend analysis, and Hurst index are utilized for analyzing the spatiotemporal evolution of shrubs in the source area. The results show that: (1) Among the three inversion models， the random forest performs the best, with R2 values of 0.91 and 0.85 for the training and testing sets respectively, RMSE values of 0.05 and 0.07, and consistency indices (d) exceeding 0.95, indicating that this model is more suitable for inverting the cold alpine shrubs in the source area. (2) The inversion results from the random forest model indicate that the area of cold alpine shrubs in the source area accounts for approximately 10.2% of the total area, mainly distributed along the valleys of the main stream and tributaries of the Yellow River within the seasonal frost zone of the source area, with a relatively high shrub coverage, while shrub distribution is sparse in the permafrost regions of the source area. (3) The results of the spatiotemporal evolution analysis show that in terms of spatial distribution, 77.93% of the shrubs in the source area are in stable zones; trend analysis indicates that from 2014 to 2024， the significantly increasing areas of shrubs in the source area account for 72.21%, while the non-significantly increasing areas account for 11.36%; results from the Hurst index combined with Sen slope analysis indicate that 85.99% of the source area is showing characteristics of sustained trends in the future. The research results can provide detailed data for monitoring spatial changes in cold alpine shrubs and understanding the response mechanisms of shrubs to climate warming and permafrost degradation.