As the world's largest CO_2 emitter,China made the commitment at the Paris Climate Change Conference in 2015 to reach carbon emissions peak by 2030,and its' CO_2 emissions intensity (CO_2 emissions per GDP) will drop by 60% to 65% based on 2005 levels. However,at present,carbon emission studies in China are mainly concentrated at the provincial and national levels. As the main role for implementing carbon reduction measures,due to the lack of basic data,cities don't have complete carbon emission inventories. To solve this problem,this paper constructed a set of city-level CO_2 emission estimation methods. The method starts with provincial energy balance tables (EBTs), and used a top-bottom approach to estimate energy consumption and carbon emissions in cities. We chose the distribution coefficient that is most suitable for the relationship between the province and city, and conducted validations for 42 prefecture-level cities' energy-related CO_2 emissions in 2012. The estimation results were compared with China High Resolution Emission Database (CHRED) and we found that the differences were all within 10%,which verified the accuracy of the method. At the same time,we studied the regional differences brought about by the method and the uncertainties that resulted from the use of different sources of fossil fuel emission factors. It is suggested that the subsequent studies should use the most appropriate local fossil fuel emission factors for city-level CO_2 emissions accounting. We believe our method can provide reasonable ideas for compiling carbon emission inventories for Chinese cities that are continuous in both time and space scales,and also can provide data support for city carbon reduction measures.