Abstarct: Geoelectrical methods are the most practical techniques in geophysics to studying the sub-surface structures. One of these methods is Resistivity Method which is applicable to investigate groundwaters, sub-surface structures and mineral exploration. Considering the significant applications of the electrical methods, inverse modeling of the measured data is remarkable. However, the geoelectrical data have a nonlinear nature and also the resulted models are not unique; and these reasons make some difficulties for the data modeling. In this research, with an extended study, an attempt has been done for solving the mentioned problems. In this way, Artificial Neural Networks (AANs) were employed for their manifold capabilities in varied branches of sciences. Thus, these networks were used in one dimensional inversion of specific electrical resistance data. In this study, a kind of network namely feed-forward neural network was used to estimate sub-surface laxyer parameters. The network is trained by artificial data which is obtained by Resix-IP. Also, it is worthy to note that the training algorithm is back propagation one. In this research, eight networks were applied to estimate the parameters of a four-laxyer resistivity model including curves types of AA, AK, HA, HK, KH, KQ, QH and QQ, and another network was used in order to classify four-laxyer models. Network response to noised data - which was added deliberately - was considered to closely resemble the artificial data with real data. Finally, the outputs of networks were compared with the results of Resix-IP with the aim of generalizing of network and acquiring precise result. The comparison indicated that the result of estimator networks response and classifying networks are outstandingly similar to the output result of Resix-IP. At last, it could be concluded that with high sufficient network structure, perfect network training as well as good produced artificial data; the feed-forward neural network not only has high precise, accuracy result, but also calculates in little time.