This study investigates the effect of soil moisture initializations and soil texture on the land surface hydrologic processes and its feedback on atmospheric fields in Taiwan. The simulations using the Weather Research and Forecasting (WRF) Model with the Noah land surface model were conducted for a 1-month period from 10 August to 12 September 2013 that included two typhoon-induced precipitation episodes and a series of clear-sky days. Soil moisture from the Global Land Data Assimilation System (GLDAS) was utilized to provide the soil moisture initialization process. In addition, updated soil textures based on field surveys in Taiwan were adopted for the WRF Model. Three WRF sensitivity runs were performed. The first simulation is the base case without any update (WRF-base), the second simulation utilizes GLDAS products to initialize the soil moisture (WRF-GLDAS), and the third simulation includes GLDAS products plus the updated soil textures and soil parameters (WRF-GSOIL). In WRF-base, the soil moisture initialization process is provided from National Centers for Environmental Prediction (NCEP) Final (FNL) Operational Global Analysis data, which are higher than the data from GLDAS products. The WRF-GLDAS and WRF-GSOIL with use of GLDAS data show lower soil moisture than WRF-base and agree better with observed data, while WRF-base shows a systematic wet bias of soil moisture throughout the simulation periods. In WRF-GSOIL, the soil textures with large-sized soil particles reveal higher soil conductivity; as a result, water drains through the soil column in a faster manner than the WRF-GLDAS, which leads to reduced soil moisture in western Taiwan. Among the three simulations, the variation of soil moisture is best simulated in WRF-GSOIL.