Utilizing remote sensing technology to scientifically assess the spatial and temporal changes of ecological environment quality and environmental response in Gansu Province is crucial for the implementation of ecological environment protection policies and the construction of beautiful Gansu. Based on the google earth engine（GEE） platform， the remote sensing ecological index （RSEI） was constructed to dynamically assess the changes of ecological environment quality in Gansu Province since the 21st century， and the Mann-Kendall and Pettitt tests were combined to determine the year of mutation. On this basis， we used ArcGIS spatial analysis， mathematical statistics， Theil-Sen Median with Mann-Kendall trend analysis， and coefficient of variation to reveal the spatial and temporal variation patterns and trends of ecological quality； we then used geodetectors and the bivariate Moran's I to identify the key drivers of the spatial variation of the ecological quality and visualize the driving effects； finally， we used the Hurst index to predict the future direction of the ecological quality. The results show that： ① The ecological environment quality in Gansu Province showed a non-significant fluctuating upward trend with the increase in years （P>0.05）， the slope of the interannual change was 0.001 3 a-1， and the sudden change node occurred in 2007. The occupied areas with excellent， good， moderate， poor， and poor ecological environment quality increased or decreased at a rate of 132.03， 1 273.44， 93.70， 1 375.66， and 63.83 km2·a-1， respectively. The spatial distribution of the quality of the ecological environment showed a polarization phenomenon， with a gradual deterioration from the southeast to the northwest. The quality of the ecological environment had a certain topographical effect， which rose and then fell with the rise in altitude and continued to rise with the rise in slope and topographical relief. ② The spatial trend of ecological environmental quality in Gansu Province was mainly upward， with the largest proportion of areas with insignificant increase. The stability of ecological environment quality was good in most regions， and the southern part of Gansu Province was the most stable. ③ Vegetation cover and precipitation were the primary drivers of spatial heterogeneity of ecological environmental quality in Gansu Province. The explanatory power was stronger after factor interaction， with vegetation cover ∩ altitude and vegetation cover ∩ temperature playing the most significant roles. The differences in spatial aggregation of ecological environment quality between the same study year and different drivers were obvious， and the spatial aggregation between different years and the same driver was highly similar. ④ It is predicted that in the future， the ecological environment quality of Gansu Province will increase in the area of 18.02×104 km2 and decrease in the area of 24.41×104 km2， and the results of the research can provide data support for the sustainable development and ecological civilization construction in Gansu Province.
Keywords: Gansu Province; Google Earth Engine（GEE）; driving factors; future prediction; remote sensing ecological index（RSEI）; trend analysis.