Based on the conventional upper-air and surface observations, temperature of brightness blackbody (TBB) from FY-2G Satellite, Doppler weather radar data and reanalysis data from ERA-Interim, two forecast failure cases of warm-sector rainstorm forecasts in severe southwest jet on 5 May 2016 and 23 April 2018 are analyzed. The results are as follows. The water vapor sources of the two warm-sector rainstorms came from strong southwest jet at 925 hPa and southwest wind respectively. The water vapor convergence in the former rainstorm had higher intensity and wider range than the latter, resulting in a broader range of torrential rain. The superposition between the strong convergence and the special topography of Nanling Mountain on the Hunan-Guangxi border could result in a stronger dynamic uplift and a longer lifetime of rainstorm. Strong atmospheric instability in the former process led to more severe rainstorms in warm sectors. The wet layers of the two rainstorms were more shallow than that in typical frontal rainstorms. And the significant wet sectors in the middle and lower layers led to the occurrence of rainstorms. The energy maintained a long time in the former process which was caused by the warm moist air mass transport of the low-level jet that led to repetitive formation of low-level convective instability with high temperature, high humidity and high energy, finally resulting in longer rainfall duration and more precipitation. The weaker moisture convergence zone ahead of the front was an important predictor of the warm-sector rainstorms in the April 2018 process. So attention must be paid to the relationship between the trough and warm-sector rainstorms. The length of meridional trough was an important factor for the precipitation zone forecast in the warm-sector rainstorms. The topography contributed to the precipitation in these two rainstorms. The rainstorm centers mainly appeared on the windward slope of valleys or basins. The rainstorms were closely related to the dynamic convergence of the boundary layer and the water vapor supply. Therefore, in forecasting operation, we should pay attention to the convergence area of the boundary layer, the special topographies and locations, and make appropriate empirical revisions to numerical forecasts so as to improve its accuracy. The radar wind profile data analysis showed that the enhancement of vertical wind shear can indicate the occurrence of rainstorms. The differences of the southwest wind thickness can reflect the difference of precipitation intensity, providing an important reference for the ascertaining of the alert level.