In this paper, fourteen MT stations along a profile perpendicular to the main geological structure were used in order to identify the geology of Sabalan geothermal field including the extension of the surface cap-rock and clay-cap layers. The MT data collected between 1-8192 Hz is of useful quality and provides good control on the surface layers in many areas. It means that reliable modeling should be possible to a depth of 1000 m depending on resistivity distribution. TE and TM-mode data and skew parameter show that the earth dimensionality differs from site to site, so we examined the 1D and 2D modeling along the profile. To have the best possible interpretation we used determinant data for 1D modeling and joint TE and TM-mode data in 2D modeling. The data was processed and modeled using an improved modeling method. One-dimensional modeling was performed using a forward operator based on reflection coefficient of the layers (Zhadanov and Keller 1998) and the two-dimensional inversion has been done by using a code from Siripunvaraporn and Egbert (2000).The resulting 1D and 2D models show a high–very low–low resistivity sequence with depth. The high resistive layer at the surface was assigned to basalt, andesitic and old trachyandesitic flows and other impermeable rocks that have thermal conduction and act as the cap-rock of the system. The role of the cap- rock is very important in sustainability of the system, because it prevents the reservoir from cooling by mixing with surface water. The second layer is a very conductive layer and is interpreted as the reservoir that has thermal convection and hot fluids contained in its fractures and pores. The resistive basement is a hot and solid magmatic intrusion and is interpreted as a heat source that produces a conductive heat flow towards the reservoir. As a result, the shallow resistivity model of the Sabalan area is in a good correlation with the geological features. It shows the common conceptual resistivity model which has been presented for the geothermal reservoirs.