Many convective storm identification algorithms based on radar data have been designed and widely used in forecasting applications not only weather departments, but also by the power departments, but the main innovation points of those algorithms are often focused on tracking and extrapolation methods. Most algorithms identify convective storm entities by defining contiguous regions that exceed one or more reflectivity threshold values. With the development of radar technology, the temporal and spatial resolution of radar data is getting higher and higher. The simple contiguous region method can’t be well adapted to higher resolution data and can’t meet diverse application requirements. This article devised a convective storm identification method based on a improved DBSCAN algorithm. Through comparing with a traditional method, it is found that the proposed method can provide more complete contour of identified storms and the identification result of high reflectivity regions is better, moreover, it has more parameters to adjust to meet more complex application requirements.