In recent years,wetland vegetation degradation has always been a global concern.It is particularly important to invert the vegetation coverage and study its temporal and spatial distribution characteristics.In order to solve the problem of mixed pixels in vegetation inversion,this paper proposes an object-oriented multi-end element spectrum hybrid analysis method.Taking Zhalong Wetland Reserve as the research object, middle-high resolution Landsat imagery is the data source,and the characteristics of spatio-temporal changes of wetland vegetation are studied from the perspective of time scale and vegetation cover level change.The characteristics of spatio-temporal changes of wetland vegetation were studied from the perspective of temporal scale and vegetation coverage level change.The results show that the object-oriented multiterminal hybrid model effectively reduces the computational complexity and the variation of the end-cells of mixed pixels,and the correlation between the inversion value and the test value is high,and the root-meansquare error is small,which is superior to the traditional multi-terminal The hybrid model method improves the accuracy of vegetation cover inversion.The vegetation coverage of Zhalong Wetland has been deteriorating for many years.The average rate of change from 2001to 2017is relatively faster than that of 1985~2000.It has important theoretical significance for improving the prediction accuracy of vegetation transfer under global climate change scenarios.