It is a great practical significance to study the green transformation of industry in Yangtze River Economic Belt. This paper take 30 provinces of China and 110 city in Yangtze River Economic Belt as sample. Firstly，study the heterogeneity of pollution discharge in the Yangtze economic belt by using the entropy weight method—Topsis evaluation model. Secondly，measure the industrial green transformation efficiency of Yangtze River Economic Belt by Super—SBM model，and analysis the dynamic evolution characteristics by the nonparametric kernel density estimation. Finally，analysis the source of power and loss. by using index of Global Malmquist—Luenberger and input—output redundancy. The results show that the pollution intensity of Yangtze River Economic Belt is nearly equal to the national average level，and the spatial difference is not significant. The efficiency of industrial green transformation in Yangtze River Economic Belt shows a trend of “convergence”，and lower than the national average level. The spatial difference between the upper，middle and lower reaches of Yangtze River Economic Belt is widened. The evolution features between the three major urban agglomeration are quite different. The driving force of the green transformation of Yangtze River Economic Belt mainly comes from the green technology efficiency and the green technology progress. The loss of efficiency is mainly due to the redundant input and undesired redundant output. The driving force and efficiency loss of industrial green transformation are different between the upper，middle and lower reaches of Yangtze River Economic Belt，so as to the three major urban agglomerations. To summarize，it is suggested to accelerate the green industrial transformation of heavily polluted cities in the upper and middle reaches of Yangtze River Economic Belt，promoting green industrial transformation according to heterogeneity of pollutant discharge，and promote the energy conservation and emissions reduction policies in Yangtze River Economic Belt based on the structural redundancy of industrial production input and output.