The CO2 emission brought about by energy, especially the fossil fuels consumption, is an inevitable by-product in the process of economic development. With the increasing accumulation of greenhouse gases (particularly the CO2), the climate has been becoming warmer. Therefore, the relationship between economic growth and CO2 emission has been paid extreme attention in the academia. Although the conventional methodology of econometrics has the advantage of testifying their correlation based on historic data, it lacks the ability of reflecting the dynamic mechanism between energy input and economic growth, which has two counteracting forces that more energy input will improve the productivity and increase the final output on one hand, but on the other hand it will also require more expenditure on energy purchase, so less remained for capital accumulation hence will deduce the final output. To deal with the weakness of econometrics methodology, this paper introduced an endogenous economic growth model. By modifying the Moon-Sonn model, we obtained the necessary condition of the existence of inverse U-shaped relationship between optimal growth rate and energy intensity, that is, the elasticity of energy in production function should be less than 0.5. Empirically, we predicted the energy intensity under present technology progress rate, which is reflected by the rate of energy efficiency improvement, and put it into the model hence it was predicted of the economic growth path with the according energy consumption under the path; Here the Input-Output analysis was applied to predict the decrease rate of energy intensity, which is 4.23% , lower than the eleventh "five-year plan" objective of 4.365%. Finally, the CO2 emission was obtained by aggregating emission from each type of energy, which requires the prediction on energy consumption structure and the estimation of CO2 emission coefficients of various types of energy. Projection results show that the peak of energy consumption and carbon emission will appear in 2043 and 2040 respectively. Furthermore, this paper also simulated the effect of energy intensity decrease rate on the peak of energy and the according per capita GDP, and constructed three scenarios to simulate the effect of reusable energy substitute policy on the peak of CO2 emission. The former simulation demonstrates a sharp influence, the faster energy intensity drops, the earlier energy consumption peaks with less per capita GDP; when the decrease rate is 4.365%, the energy will peak in 2041, 2 years earlier than the reference with the per capita GDP is 125621 Yuan RMB; when the decrease rate is between 4.5%-5%, the peak year of energy consumption will advance to 2040 or even earlier when the per capita GDP will approximately be 100000 Yuan RMB, much approach the situation of OECD member countries. And the scenario analyses show a dramatic decline of emission volume when the proportion of renewable energy rises, but the reusable energy substitute policy have an insignificant influence on the year when the peak of CO2 emission shows up.