Abstract: To address the problematic time weights of panel data in the evaluation of carbon emission efficiency, a partial-order set evaluation model was used to evaluate the carbon emission efficiency of China. In this model, it was not necessary to know the specific weight values of the indicators, but only the weight order was needed. By collecting the panel data of 30 Chinese provinces (autonomous regions and municipalities) from 2000 to 2017, calculating each year′s carbon emission efficiency, and sequencing the weight values in time reverse order, the carbon emission efficiency was evaluated. The results showed that the carbon emission efficiency was the highest in East China, followed by the Central China, and the worst in Northwest. The factors affecting the carbon emission efficiency in each region were different, which required that specific carbon emission policies should be developed according to local conditions. The carbon emission efficiency of each region tended to be aggregated, and the local carbon emission policy would not only affect the carbon emission effect of the region, but also affect the carbon emission of the adjacent areas, so, it was necessary to strengthen the cooperation of carbon emission reduction among all the regions.