Mountain rangelands play a key role in providing forage, preserving biodiversity, and regulating the carbon cycle. The aim of this study was to predict the long-term grazing capacity of the summer rangelands in the Arshad Chaman Kandovan Sahand region and to investigate the impact of climate change on vegetation production and cover. Field data were collected over a 16-year period from 2006 to 2021 along six transects. Vegetation indices, including cover percentage and forage production in three palatability classes (I, II, and III), as well as soil surface conditions, were measured. Climatic data, including minimum, maximum, and mean temperature and precipitation, were obtained from Tabriz, Sahand, and Maragheh stations and analyzed using SPI and SPEI drought indices. General Circulation Model data were downscaled using the SDSM model under RCP scenarios, and XGBoost, Random Forest, and Linear Regression machine learning models were used to simulate grazing capacity. Results showed that rangeland forage production decreased from 1250 kg/ha in 2006 to 1054 kg/ha in 2021, representing a 15.6% reduction, and available forage decreased from 408.2 to 316.7 kg/ha, representing a 22.4% reduction. Consequently, short-term grazing capacity declined from 7.3 to 5.7 Animal Unit Months per hectare. In grazing capacity prediction, the XGBoost model, with an R² of 0.96, showed a 3.5% reduction by 2051, while the Random Forest model, with an R² of 0.85, showed a 3% reduction. Based on results, it is recommended that livestock numbers be reduced by approximately 20% by 2051. Vegetation enrichment programs through rangeland improvement practices, such as seeding and grazing management, should be implemented with local community participation, and continuous monitoring should be maintained.