This study applied the dynamic hydrologic model (TRIGRS), the steady-state critical rainfall value (Qcr), and the critical rainfall line (CL) to estimate the critical rainfall for regional shallow landslides. Using the rainfall-induced landslide case of the BangFu North Creek sub-watershed in the KaoPing River watershed during Typhoon Morakot in 2009 as an example, the three methods were compared to discuss their differences and adaptabilities. The analysis results of the dynamic model (TRIGRS) show that the rainfall threshold, defined as the cumulative rainfall at a 50% cumulative area ratio of the landslides in the study area, is around 328 mm. Furthermore, the results of steady-state modeling show that the critical rainfall values (Qcr) vary spatially from 50 to 400 mm/day. Finally, we attempt to plot the result of dynamic model analysis on a rainfall intensity- duration graph and obtain the landslide critical line for the study area. The slope of the critical line is similar to other data from abroad but the rainfall intensity at the landslide occurrence for the study area is much higher than the experience abroad. The preliminary results could be regarded as a good reference for authorities responsible for prevention and mitigation of landslide disasters.