The calculation process of Hurst index needs to group the data of time sequence. And different grouping situations have different influences on the Hurst index. Take daily closing level of the Index from 1997 to 2001 in Shenzhen stock market for example, it shows that the calculated Hurst index has more reliability when there are about 15 data in a group. But different grouping methods will not change the quality of conclusion but only its quantity.