The Martian magnetic field, characterized by its complex and heterogeneous structure, poses significant challenges for modeling due to the inaccessibility of the planet's interior and noisy satellite data. This study presents an iterative approach using the S-approximation method, a robust analytical method for handling large but fragmented data sets from orbital missions on Mars. By representing the magnetic field as a sum of fields generated by simple and double layers on predefined surfaces, this method allows for the construction of high-resolution models that reflect the observed magnetic field. The approach allows for flexible selection of carrier geometries (planes, dihedral angles, spheres, ellipsoids) to accommodate different study scales and problem geometries, facilitating the modeling of local and global magnetic anomalies. By choosing different model parameters and integrating different data sets, this approach improves the accuracy and spatial coverage of Martian magnetic field models.