Commodity futures trading plays a very important role in China's current economic system. Investors all hope to get some profits from a large number of futures trading. However, as a kind of speculation, traders often have to bear great risks in it. This paper studies some problems in commodity futures trading, and gives trading methods to obtain greater profits. Question 1: firstly, we use the model prediction method in SPSS to give the fluctuation chart of each index of rubber futures trading on September 3, and then use Matlab software to make the correlation chart between the trading price and each index. The conclusion of the analysis is that the transaction price of commodity futures is significantly related to the prices of B 1 and S 1, and also related to the increase and decrease of trading volume and positions, the number of B 1 and the number of S 1, but not related to the total amount. Finally, using SPSS software to carry out binary correlation analysis to further determine the relevant indicators. In order to classify these changing characteristics of rubber futures prices, we made a trading price fluctuation chart of 19, and analyzed the changing characteristics of other indicators by taking positions as an example, and divided the seven indicators into two categories: rising and periodic fluctuations.

Question 2: In this paper, the method of regression analysis is used to establish the price fluctuation prediction model. Firstly, the basic principle and content of regression analysis are introduced, and the least square method used in regression analysis is expounded. Then on the basis of the first question, the mathematical model of regression analysis is established, the functional relationship is obtained, the fluctuation trend of price is calculated and compared with the actual data. Then, the residual data in the model are analyzed to verify the rationality of the established regression model.

Question 3: In order to establish the trading model of maximizing returns, we analyze the fluctuation data of prices, and then analyze the "high" and "low" prices with the help of the moving average theory, and get some selling points. After the trading model is established, the appropriate trading opportunity is analyzed by MATLAB software, and the graph is drawn, and the income is calculated according to the established model by using the given data.