This dataset focuses on 282 prefecture-level cities in China, with a time span of 2004~2020. Core variables include green economic efficiency (gee), population agglomeration (pop), economic agglomeration (eco), industrial agglomeration (ind), as well as agglomeration indicators with alternative measurement methods (npop, neco, nind). It also covers influencing factors such as urban infrastructure (infra), government attention (gover), opening-up (open), instrumental variables like railway accessibility (rail) and terrain undulation (tu), and classification information of region (area), city size (size), and province (pro). Missing data is filled with the mean of adjacent values, with a data volume of 364.17KB in a single file. It is openly available under the CC0 license and can support research on the relationship between urban agglomeration mechanisms and green economic efficiency.