The Price Prediction Model Of The Futures Cotton Between CZCE And NYEX

China's futures market has been set up more than 10 years. More and more enterprises, individuals and institutional investors have joined the futures market. The original intention of establishing a futures market was to avoid the risk of spot market, and judge futures prices trends. Cotton futures have become one of focus Commodities in recent years. On one hand, China is the largest cotton production and consumption country. The China factors affect world cotton prices trend. On the other hand with the increasing textile exportation, textile has become China's great favorable trade balance of goods. The trade dispute on textiles become. And cotton is one of the most important raw materials of textiles.In this paper, it is primary to test the futures copper prices of China Zhengzhou Commodity Exchange (CZCE), New York Board of Trade(NYBOT) ,Commodity Research Bureau Futures Price Index(CRB)and China Cotton Index stationary at levels. After demonstrating the Granger causal relationship between those series, a VAR model is built. The result shows that futures copper prices series of China Zhengzhou Commodity Exchange follow the VAR model. Moreover, our empirical analysis shows that the changing of futures cotton price of CZCE can show much information to Cotton Grower and Textile Goods enterprise.The main feature in analyzing cotton futures prices is the use of modern econometric theory. Particularly through the use of unit root test and co-integration theory discuss the cotton futures price formation mechanisms. And after Granger causality test and Johansen co-integration test, VAR model via those time series. In the simulation results can be seen that the VAR models can effectively capture the range and trends of cotton price changes. But in all honesty, this paper still needs further strengthening. This is due to China's futures market and cotton futures trade had not long histories. The volume of trade data collection is limited. And as limited time interaction analysis between every two series failed to carry out.Most of the statistical modeling of text processing is done by econometric softwareEviews.