Using the instability index computed with radiosonde, NCEP FNL data and hyper spectral Atmospheric Infrared Sounding (AIRS) standard retrieval data, this paper analyzes the atmospheric instability before occurrence of the severe convective weather in Beijing on June 23, 2011. The findings suggest that the instability index computed with radiosonde shows the change of the unstable energy over Beijing before and after “the 23 June Storm”, and no sounding data in the upstream key region. The instability index computed with AIRS and NCEP shows that upstream key region in Beijing is extremely unstable before the occurrence of the severe convective weather (K index is greater than 40, and SI index is equal to or less than -5), being conducive to the occurrence of severe convective weather. The research results of the paper show that the radiosonde data cannot monitor the severe convective weather because of its lower spatial and temporal resolution. After quality control the instability index calculated by AIRS data can be used in monitoring the occurrence of convective weather. The NCEP data with lower spatial resolution have weak capability in monitoring small scale atmospheric unstable stratification. In summary, AIRS retrieval products have the advantage of high spatial and temporal resolution. We can monitor the atmosphere unstable energy accumulation of the upstream key areas before “the 23 June Storm” weather occurs by using the clear atmospheric instability index computed with AIRS L2 products, providing supplementary information for forecasters.