Through exploring price characteristics of carbon futures products in EU ET, this paper aims to provide China's policy makers with meaningful materials and references for understanding how a carbon trading market can be established and well regulated. Based on the dataset comprising of multiple sources including Euro stoxx600 index, coal and crude oil prices, natural gas prices and European clean energy company stock prices, etc., this paper uses BP neural network model to simulate the long-term trends of carbon futures prices in six scenarios that represent the typical features of a carbon trading market. The results show that: (1) the magnitude of economic development's effect on carbon price is the largest among other factors, with the shortest duration; (2) in comparison, the effect of black energy consumption is weaker, but its lasting duration is the longest; (3) the impact of clean energy development on carbon price is similar to that of black energy, but the effect magnitude and lasting duration are relatively smaller. These findings suggest three viable directions for the development of China's carbon trading market in future i.e. adjusting total quotas in accordance with economic development, establishing market price stabilization mechanism, and developing clean energy. The novelty of this paper is to simulate the long-term trend of carbon prices by constructing a carbon price prediction system.