Abstract: Based on the data of routine observations,regional automatic stations and Himawari-8meteorological satellite,the 43 severe convection weather processes that occurred in southeast Gansu between April and September 2016 were analyzed.The indicator of severe convective cloud recognition,tracking method and forecast indicator were established,and some individual cases of 2018 were tested to validate the effectiveness.The results showed that:（1）Using double threshold values of TBB≤238 K of the B13 channel（10.4 μm）or the TBB difference △TBB＜0 K of the B08 channel（6.2 μm）and the B13 channel as the indicator of severe convective cloud recognition can accurately identify severe convection weather clouds in southeast Gansu.（2）Approaches like“reverse searching”and “area overlap”could be used to calculate the center of gravity of clouds,thereby accurately locating and tracking severe convective clouds and extrapolatively forecasting their paths of movement.（3）The indicator to discriminate the fallout zone of convective weather has a certain forecast capacity for the fallout zones of short-time strong rainfall and hail within the region.