Abstract: Based on the radar mosaic 3D data,automatic weather stations observations and disaster wind data, twenty cases of thunderstorm gale from 2008 to 2012 in the Beijing-Tianjin-Hebei region are statistically analyzed to develop an automated detection of thunderstorm gale with the fuzzy logical based algorithm. The capability of the algorithm is examined. In the method, the six main radar identification indices of ground gale are given with the corresponding membership functions and weight coefficients determined. All the gale is tested and analyzed, including the three types of echo, i.e. the massive echo, banding echo and floccus echo. The results show that the massive echo is triggered by the strong storm monomer with strong echo, higher echo top, bigger vertical integrated liquid water content (VIL) value and faster moving speed, and in this case the gale occurs nearby the thunderstorm cell with the same route as that of the cell; the banding echo mainly contains the squall line and the bow echo with its length greater than the width and the impact range of strong wind located at the forefront of the band echo; the floccus echo generally refers to the mixed echo with large area layer echo embedded by the isolated massive echo, and the strong wind area situated around the storm monomer. The wind range for the three types identified is generally consistent with the real wind, and the hit rate of massive echo, banding echo and floccus echo is respectively 96.2%, 68.6% and 45.3% with the missing rate of respectively 3.8%, 31.4% and 54.7%. Lower omission rate of floccus echos gale is because of weak echo intensity and lower VIL and missing rate is caused by the sparse station distribution and algorithm. This also proved that the automatic identification method is efficiently and feasible, it has important practical guiding significance for an operational system in short-term forecasting and nowcast warning. The work also provides a foundation in warning the position of surface gale.