Improving green economic efficiency (GEE) and productivity are crucial for China to realize sustainable development goals. However, the economic growth of China has followed an extensive development model with high energy consumption and heavy pollution. This study conducts data envelopment analysis (DEA) to evaluate the GEE of China. First, we introduce the polar coordinates theory in the epsilon-based measure (EBM) model to construct a Ray epsilon-based measure (REBM) model. In addition to the merits of EBM model, the REBM model accounts for the weak disposable relationship between undesirable and desirable outputs. Second, based on REBM model, a REBM-Malmquist-Luenberger (REBM-ML) index is constructed to evaluate the green total factor productivity (GTFP). Finally, we conduct spatial econometric analysis to reveal the dynamic evolution of GTFP. According to the empirical results, the GEE of China is generally low, and the urban agglomerations located at coastal regions own higher GEE. However, the GTFP made progress overall, mainly benefited from the technical progress. Accordingly, in the process of sustainable development, China still faces the challenge of energy saving and emission reduction. The spatial econometric analysis reveals that the GTFP of China existed a significant divergence trend and there was spatial spillover effect between cities, as well as urban agglomerations. Furthermore, we provide policy implications and suggestions for Chinese sustainable development.