Based on the data of industrial added value growth in China from January 1991 to December 2008, the ARFIMA-FIGARCH model is used to measure the dynamic change process of the real output growth rate. It is found that the first order moment of the growth rate of China's industrial added value does not exist Long-term memory, and its second-order moments have significant long-term memory, which means that the Chinese output growth rate level sequence does not have long-term memory characteristics, while the output growth uncertainty sequence shows long-term memory; The Granger influence relationship between the output growth rate and its uncertainty based on the VAR model shows that the output growth rate has a significant one-way Granger influence on its uncertainty. Therefore, when the economic policy is selected , The characteristics of the relationship between output growth and output growth uncertainty should be fully considered.