Green development is the core manifestation of the quality and competitiveness of urban development. Based on the perspective of efficiency, this paper uses urban panel data and unexpected output SBM model, measures the level of urban green development in China, analyzes their temporal evolution characteristics, and uses coefficient of variation and double fixed effect Spatial Dubin Model to test the convergence of urban green development in China. The research findings are as follows: Firstly, during the investigation period, the level of urban green development in China is relatively low, showing an overall fluctuating upward trend. Secondly, there are obvious regional differences in the level and growth rate of urban green development in China. Thirdly, based on the common characteristics of changes at the national and four regional levels, the level of urban green development in China shows short-term σ convergence and divergence characteristics only in 2004—2006 and 2017—2019, while in other years, there is no convergence or divergence trend. Fourthly, the spatial convergence analysis shows that there is no absolute β convergence or conditional β convergence characteristics in the level of urban green development, whether it is the national city sample or the four major regional city samples.