Abstract: Natural gas is an important resource, and it has many advantages such as high efficiency, cleanliness and safety.Nowadays, it has been widely used.Accurately forecasting the demand for natural gas in future will have a significant positive effect on the formulation of energy strategy and on the production, import and export of natural gas.In order to predict the demand of natural gas from China, a new hybrid optimization method, DDE-BAG, is proposed in this paper.This method is based on the original DEBA algorithm and uses the adaptive principle to adjust the mutation operator and the crossover operator, change the way of disturbance and mutation obtained.In this paper, we chose four factors of living gas consumption per capita, the proportion of natural gas consumption in energy consumption, economic growth (GDP) and population urbanization rate as the input factors.Using data from 1986 to 2015 for 30 years as the observation data, the DDE-BAG algorithm was used to optimize the coefficients of multiple linear and exponential forms of natural gas demand estimation models, and the optimal coefficients of the two models were obtained.The results showed that the predicted values and observed values of the two models were close to each other.Finally, two models were used to analyze the predicted values of two models in China from 2016 to 2030 during the forecast of natural gas demand.The results showed that, even though the forecasting methods adopted were different, the results of the forecasting were similar.