It is generally thought that vegetation has a stabilization effect on a slope; however, very limited quantitative field data are available for verifying this perception. In order to improve our understanding on the vegetation effect on rainfall infiltration and hence on slope stability, a well-instrumented field study was carried out on an unsaturated expansive soil slope in China. The field program consisted of two neighboring monitoring areas (both 16 m wide by about 30 m long): namely a bare area and a naturally grassed area (real slope). Artificial rainfall events were produced in succession in the two areas with a specially-designed sprinkler system. In this paper, the relevant monitored results from the two areas are directly compared and discussed. Prior to the artificial rainfall, the measured higher initial soil suction in the grassed area than that in the bare areas was attributed to the evapo-transpiration effect of the grass. During the rainfall, the presence of the grass greatly increased the infiltrability of the upper soil layer and delayed the onset of surface runoff. The observed delayed responses (i.e., about 3 days) of surface runoff, soil suction and water content to the rainfall in the grassed area was about twice the duration of the delayed response in the bare area (i.e., about 1.5 days). The influence of the simulated rainfall on the changes of soil suction and water content in the grassed area was found to be more significant and about 1.5 m deeper than that in the bare area. The greater depth of influence observed in the grassed area may be attributable to the greater depth of open cracks due to the evapo-transpiration effect of grass.