[Objective] The spatiotemporal patterns of rainfall-induced landslides in Chongqing City and their correlation with rainfall were analyzed， in order to provide a scientific basis for regional early warning and prediction of such landslides， with the goal of reducing losses caused by geological disasters. [Methods] Using rainfall-induced landslides during the 2024 flood season in Chongqing City as the research subject， statistical and geospatial analysis methods were applied to examine the geological environmental characteristics of landslides and their association with rainfall. I-D， E-D， and E-I rainfall threshold models were established， and their accuracy was compared using a confusion matrix. The predictive capability of the models was evaluated using landslide data from June 2025. [Results] ① Landslides in the study area primarily occurred in low-elevation areas （500—1 000 m）， on north-facing slopes， with gradients of 10°—30°， in soft rock formations， and within 800 m of folds and river systems. ② The occurrence of landslides was closely related to rainfall， exhibiting clear critical threshold effects， hysteresis， and nonlinear characteristics. ③ Among the various rainfall threshold models established， the I-D model demonstrated superior comprehensive predictive performance compared to others， making it more suitable for landslide early warning in the region. ④ The established I-D threshold model for Chongqing’s flood season landslides was region-and period-specific， with values higher than global and annual thresholds for Chongqing but lower than thresholds for southeastern China and national extreme events. It can serve as a key criterion for early warning during the flood season in this area. [Conclusion] The geological environment controls the spatial distribution pattern of landslides， while rainfall is the key triggering factor. The established I-D threshold model provides a critical basis for regional early warning. Achieving precise early warning in the future will require moving beyond single rainfall criteria and developing dynamic comprehensive models that integrate multiple factors.