Abstract: Preliminary theoretical studies were dealt with in this paper on the efficiency of one-dimensional inversion of apparent resistivity data observed with multi-separation array, for understanding the true resistivity variations with time inside the media. To this end a three-layer model has been taken as an example. Firstly, we did the inversion of the theoretical apparent resistivity data by forward modeling calculation for each layer in one-dimensional resistivity structure. The result shows a good consistency between the true resistivity variations and that obtained with inversion of each layer, suggesting that one-dimensional inversion technology is adequate for recognizing the variations of resistivity in each layer. Secondly, more complex model was taken due to some interface fluctuation, which deviated from one-dimensional model to some extent. The same process of forward-inversion was conducted in the case of the fluctuation of interfaces of the three-layer model, the results showed that good fitting can be found only in the condition of non-large variations of resistivity of the first-and second-layers, otherwise the inverted resistivity of each layer will turn up distortion. Usually the shorter the time interval of observation, the smaller the medium resistivity change during the two consecutive observation. Therefore, to reduce the distortion, it could be available by taking short time interval of observation with multi-separation array.