This paper analyzes the variation of cotton spot,futures and e-market price in China by using Johnsen co-integration test,Granger Causality test,VECM model,Impulse Response Function and Variance Decomposition method. The result shows that three market prices demonstrate more consistent fluctuant trend,and maintain a relatively stable equilibrium relationship for a long term. There is significant bidirectional Granger causal relationship among them,however,the dominant levels are not the same. Cotton futures price has a dominant position over cotton spot and e-market,on the contrary,their influence on futures is much smaller. The impact of cotton spot's influence on futures is much larger than that of futures in e-market. The impact of cotton spot's influence on e-market is larger than that of cotton spot in e-market.