This paper is the result of ecological geological survey engineering. Objective Vegetation cover is a critical indicator of ecosystem stability. Researches about the factors influencing vegetation cover changes in the ecologically sensitive Kunlun Mountains region are limited. This study aims to analyze the spatiotemporal variations of vegetation cover and their driving factors in the Kunlun Mountains and to perform a regional ecological assessment. Methods The Normalized Difference Vegetation Index (NDVI) was used to examine the spatiotemporal patterns of vegetation cover from 2000 to 2020, and the Geodetector (GD) and Random Forest (RF) models were applied to identify the primary drivers of NDVI changes. Meanwhile, RF and Long Short–Term Memory Network (LSTM) models were used for predict the NDVI variations, and the Remote Sensing Ecological Index (RSEI) was used to evaluate environmental quality. Results NDVI showed a generally temporal increasing–decreasing–increasing trend, characterized by a spatial distribution that higher values in the west and north and lower values in the east and south. The RF model identified the digital elevation model (DEM), precipitation (Pre), distance to towns (Town), and gross domestic product (GDP) as the most influential factors, whereas the GD model showed evapotranspiration (ET), DEM, GDP, and temperature (Temp) had the greatest explanatory power. The RF model was more effective for NDVI regression analysis than the LSTM model in the study area. The RSEI value showed an increase in 80.52% of the area, but the overall variation was considerable, with 78.86% of the area showing significant variation. Conclusions DEM, GDP, Pre, and ET are the primary factors affecting vegetation cover in the region and the significant improvement in the regional ecosystem over the past two decades. It also reveals a significant enhancement in regional ecological quality, indicating that ecological conservation measures have been effective.