Taking two large runoff events and three small runoff events in the Qingjiang River Basin from 2016 to 2017 as the research objects, we analyze the spatial and temporal distributions of precipitation from the three products of CMORPH satellite-gauge-radar merged precipitation product (CMPAS), China global atmospheric reanalysis product (CRA), and rain gauge precipitation (Gauge). Then, based on the characteristics of runoff events and different precipitation products, the two experiments are designed, and the runoff simulations of WRF-Hydro model driven by the precipitation data from the three precipitation products are analyzed. Finally, combined with the spatiotemporal variations of precipitation, we explore the effectiveness of the three precipitation products in the runoff simulations. The main results are as follow. (1) The rainfall centers and the rainband locations and orientations detected with the three precipitation products are similar for the five runoff events, and the temporal trends of rainfall averaged over the Qingjiang River Basin are consistent. (2) All three precipitation products can produce the runoff events in both experiments. For large runoff events, CMPAS performs best, with correlation coefficients above 0.76 and Nash efficiency coefficients above 0.63. For small runoff events, Gauge performs best, with correlation coefficients above 0.75 and Nash efficiency coefficients above 0.48. CRA has a relatively poor performance for both large and small runoff events, but its simulations are significantly improved after calibrating parameters. (3) After calibrating the parameters in the second experiment, the simulation results of runoff are improved to different extents for the peaking, rising, and falling periods of each event. For the small runoff events, compared with the rising and falling periods, the simulations driven by each precipitation product for the peaking period is significantly improved, while, for the large runoff events, the simulation results driven by the three precipitation products for each period is not significantly improved.