Abstract: Spin-up process is defined as the adjustment process of a model before it reaches an equilibrium state under the circumstance of given initial non-equilibrium or disturbed conditions. Based on the reanalysis data from NCEP (National Centers for Environmental Prediction) and the observation data of surface energy fluxes and soil temperature at Zhangye station in the HEIFE experiment conducted from 20 June to 20 July 1991, we studied the characteristics and time range of spin-up processes in the land surface model of RAMS (Regional Atmospheric Modeling System). To estimate the influence of initial soil moisture and temperature on spin-up process, we designed more than 40 sets of initial soil parameters. After analyzing simulation results, we conclude that there are four kinds of typical spin-up behaviors, and the time required for the model to reach an equilibrium state varies greatly with different initial soil parameters, ranging from several hours to more than one week. Soil moisture content is a very important factor in estimating whether the land surface model has reached an equilibrium state or not. Our conclusion is useful for model runners to judge whether their model has reached an equilibrium state, and thus beneficial to improve the accuracy of meso-scale model.