Using the equation for magnitude of the inferred largest aftershock proposed by Shcherbakov and Turcotte(2004), we estimate the magnitude of the largest aftershock following the Jiuzhaigu earthquake from catalogues with different time intervals after the mainshock. The results show that the estimation trends to be a certain value when time intervals become larger and this value is consistent with the magnitude of the real largest aftershock. It should be emphasized that if the aftershock data are relatively complete, using the data in short time intervals after the mainshock(1 day or 2 days)can give an accurate estimation of magnitude of the largest aftershock of the Jiuzhaigu earthquake catalogue. In fact, estimations with different time intervals after the mainshock also indicate that using the aftershock data measured 12 hours(0.5 days)after mainshock can give a rational lower bound of magnitude of the largest aftershock. Moreover, we use two types of aftershock catalogues following the Jiuzhaigu earthquake with Richter magnitude(M_L)and surface-wave magnitude(M_S). And the corresponding results show that the two different catalogues give consistent results which indicates that magnitude of the inferred largest aftershock doesn't depend on types of magnitude chosen for catalogues. Therefore, the method we proposed can be generalized and applied to seismic hazard analysis in southwesten China. Current fitting method also implies that with the help of advance in earthquake detection and aftershocks identification, the method we proposed can accurately predict the magnitude of the largest aftershock in a short time after mainshock. In a word, although we can't predict precisely when and where the largest aftershock will occur, the method we proposed can provide rational magnitude estimation for the largest aftershock in a short period after the mainshock.