Under the background of responding to global climate change, China is facing huge pressure of reducing carbon emission at present and in the future, either for its own economic development needs or abiding by international conventions. Accurately estimating China s carbon emission is vitally important to make polices for reducing carbon emission and dealing with international negotiations. Although there are many studies on thesis of estimating China s carbon emission, the estimations on China ' s carbon emission varied remarkably among different studies. Based on the broad and intensive literature reviews, the paper adopts a meta-analysis method and the multi-factor variance analysis (ANOVA) to analyze the key factors that cause the large variations of estimations. Our results indicate the sources of researches and the choice of carbon emission coefficients affect the estimation significantly. The sources of researches, choice of carbon emission coefficient, energies classification and calculation based on national or provincial data are the four key determinants, accounting for 30.5%, 29.7%, 21.2% and 11.8% of variances of estimations respectively. According to our findings, it is proposed that the future studies to estimate China s carbon emission should pay extreme attention on those key elements, and it should adhere to research output of Chinese authorities and scholars in international negotiations.