Abstarct: Source parameter estimation techniques are among the interpretation methods of potential field (gravity and magnetic) anomalies. The source is the subsurface causative body having physical parameters such as boundary, depth, dip, remanent magnetization and so on. In practice, estimation of the depth, boundary and remanent magnetization of the body are more important then the other parameters. Nowadays, Euler deconvolution, analytic signal and wavenumber methods are extensively used in ground and airborne magnetic and gravity surveys for estimation of the source parameters. The main aim of this research is to improve the effectiveness and performance of these three methods. In this research, first, the mathematical behavior of the analytic signal baxsed on harmonic and homogeneity analysis was investigated. baxsed on this analysis, it was found that the analytic signal is a homogeneous and non-harmonic function. Thus, we cannot compute its vertical derivative in frequency domain. For this, an algorithm baxsed on the finite difference method in the space domain was introduced. The numerical analysis of this algorithm was accomplished in MATLAB software. In addition, a method baxsed on the multiples of analytic signal was used for depth estimation of 2-D bodies. The advantage of this method is that it does not need analytic signal derivative so its sensitivity to noise enhancement is minimum. This method, after testing on synthetic data, was applied on magnetic data from Soork iron ore mine in Yazd Province and the obtained results were compared with drilling data. Moreover, magnetic analytic signal tensor using Hilbert transform was defined, and also, a new edge detection filter called BD was introduced. In order to estimate the depth and structural index of the magnetic body simultaneously, the analytic signal tensor (AST) with Euler deconvolution was combined leading to a linear equation that was solved using least square method. This method was applied on magnetic gradient data from Siriz and Soork iron ore mines, located respectively, in Kerman and Yazd Provinces, and the obtained iron ore depth results were validated with drilling data. Findings show that this method is noise sensitive and for more accuracy, upward continuation or measured gradient data have to be used. The average estimated depth of iron ore of 57.6 meters has been obtained from this method while the average depth of iron ore in drillings has been 69.4 meters. The precision and accuracy of the gradient-baxsed methods are dependent on selected geometry of the body. However, total-baxsed methods such as SLUTH are body independent and have multi-scale nature. This methods use measured total magnetic field data. The results of the AST-Euler combination method were compared with the results of the SLUTH method. The algorithm of the SLUTH method was modified so that implementation of the upward and RTP filters as well as estimation of the structural index and depth simultaneously became possible. This method after testing on synthetic model data was applied on Soork magnetic data, and the obtained results showed good agreement with the drilling results. Wavenumber as a complex attribute is a recently used function in regional magnetic interpretation. The complex attributes due to a body with polygonal cross section was introduced. Furthermore, the tilt angle of a dipping contact model with arbitrary magnetization direction was used to estimate dip, trace and depth of the fault model by means of equivalent source principal. For this purpose, a MATLAB code was developed and tested on synthetic and real magnetic data. This method was applied on airborne magnetic data from a fault in the United States of America and also on ground magnetic data from Anarak area of Isfahan Province