The main goal of the present study is to identify characteristics of the inter-station horizontal magnetic responses and the vertical magnetic data, as two types of magnetotelluric transfer functions, in the modeling procedure. Through consideration of model responses and two-dimensional inversion of synthetic data, sensitivity of the data components in detecting different geophysical structures is investigated. An inversion of the magnetic transfer functions related to a real data set has been also accomplished, and the importance of a reference selection in analysis of the inter-station horizontal magnetic data has proved. Analysis of the results of the synthetic tests confirms the efficiency of the modeling based on horizontal magnetic responses and indicates more accuracy of this type of data, especially in terms of the resolution of the deep targets. There is not a great difference in the resolution of the structures between two datasets, as closely located anomalies are examined. Furthermore, both transfer functions can also sense vertical conductivity distribution.