Based on panel data of 30 provinces in China from 2005 to 2021,this paper firstly estimates the agricultural ecological value scientifically,then incorporates it as a desirable output into the total factor productivity measurement system,and constructs the EBM-GML model to measure and decompose the growth of agricultural ecological total factor productivity in China.It also analyzes the spatial-temporal evolution characteristics of agricultural ecological total factor productivity growth by using Dagum's Gini coefficient,convergence model,kernel density estimation and Maikov chain.The findings are as follows.Firstly,from 2005 to 2021,China's agricultural ecological value showed a fluctuating upward trend and the spatial distribution pattern was"the highest in the west,lower in the middle,and the lowest in the east".Secondly,the average annual growth rate of China's agricultural ecological total factor productivity was 3.08%,showing periodic fluctuating growth and obvious spatial non-equilibrium characteristics.Its growth was mainly driven by ecological technology progress.Thirdly,the overall difference in the growth of agricultural ecological total factor productivity showed a significant expansion trend,mainly due to inter-regional differences.In addition,there was no σ convergence and unconditional β convergence for both the whole country and the three major agricultural regions,but there is conditional β convergence.Lastly,the regional gap in the growth of agricultural ecological total factor productivity had widened significantly,and the distribution pattern had a tendency to evolve from"single polarization"to"polarization"and"multi-polarization",and the internal mobility of distribution dynamics was weak