Soil moisture is of critical importance to the physical processes governing energy and water exchanges at the land-air boundary. Since the soil moisture reservoir has a memory considerably longer than that of most of the atmospheric processes, a climatic anomaly may persist through processes dependent on soil moisture. Observations have provided a description of the soil moisture profile variability phase shift, fluctuation damping and persistence increasing with soil depth. This variability as a function of soil depth is linked to the time scales of the atmosphere and the climate spectra. Whether or not climate models can reproduce this variability should be a good test of their land process representations in the treatment of soil hydrology. Meanwhile, it could be an effective physical approach to improve the accuracy of the atmospheric prediction, especially for the lower frequencies. A multi-layer land surface model is applied for simulating the soil moisture profile variability observed from observations, and for exploring the underlying mechanisms. The sensitivity experiments are conducted by examining the impacts of various potential controlling factors, (i) the initial soil moisture field, (ii) the atmospheric forcing at the upper boundary, (iii) the root sink term related with vegetation type, and (iv) the soil texture, on the soil moisture profile variability characteristics, that is the phase shift and amplitude damping. The simulated key land surface prognostic variables, soil moisture and evapotranspiration, are evaluated against observations prior to the sensitivity study.