Abstract: Using rainfall data from 14 national meteorological stations and 466 regional automatic weather stations in Chengdu, together with MICAPS and ERA5 reanalysis data from 2004 to 2023, this study analyzes the spatiotemporal distribution as well as interannual and monthly variations of regional rainstorms in Chengdu. By integrating topographic features, this study classifies the circulation patterns associated with rainstorms, and establishes the corresponding conceptual models and forecasting focus. The results show that: (1) Rainstorm frequency exhibits strong topographic dependence, decreasing from west to east. Western-type rainstorms occur 2.5 times per year on average, central-type 1.8 times per year, and eastern-type fewer than one time per year. (2) A total of 111 regional rainstorms occurred in Chengdu over the past 20 years, peaking in 2013 and 2018 (12 events respectively), with most events concentrated in July and August. (3) Western-type rainstorms account for 49% of the total, central-type for 35%, and eastern-type for only 16%, less than one-fifth of the total. (4) Regional rainstorms are classified into five types: the east-high and west-low type, the two-trough and one-ridge type, the subtropical high-edge type, the two-high shear line type, and the northwest flow type. Among them, the east-high and west-low type occurs most frequently, accounting for 44%, followed by the two-trough and one-ridge type, and then the subtropical high edge type. The two-high shear line type and the northwest flow type have the lowest occurrence frequency.(5) Based on the vertical configuration characteristics at different levels, including the positions of 500 hPa troughs, ridges and low vortices, the orientation of the 700 hPa low-level shear line, the intensity of the southwesterly jet, the southwest vortex activity, moisture transport axes, as well as the 850 hPa wind convergence zone, high-humidity zones and topographic effects, this study summarizes multi-system integrated conceptual models and clarifies corresponding forecasting focuses. These models provide key diagnostic indicators for rainstorm spatial forecasting and offer valuable references for operational prediction of regional rainstorms in Chengdu.