Aiming at solving the problem of wind disaster information acquisition in Guangzhou, Guangdong Province, this study established an application model based on the characteristic analysis of location microblogs. The acquisition model has such functions as semantic retrieval and filtering analysis, disaster information classification and mapping, disaster situation weighting and GIS hotspot analysis. The model was applied to extract the wind information in Guangzhou, during the strong convection weather in 21 April 2017. The result shows that the wind information acquisition model has a certain practical value for wind information acquisition problems. Compared with the conventional method of wind information acquisition, the location microblog data was more rapid, low cost and abundant. This model provides a new method for wind disaster information extraction.