Abstarct: Through being a lot of sedimentary structures in Iran such as oil traps which are economically important, study and modeling of baxsement topography might be deliver valuable information. To goal a thorough interpretation of baxsement topography modeling might be used. In this study, 3D non-linear inverse modeling of gravity and magnetic data is used to determine topography of baxsement. Two most useful models for modeling a three-dimensional baxsement are juxtaposing rectangular prisms (Plouff, 1976) and the polygon (Talwani and Ewing, 1960). With the model of juxtaposing rectangular prisms, inversion become not only faster, also will be avoid of the difficulties associated with the model of Talwani and Ewing (Rao et al, 1999). In this study, a three-dimensional baxsement is modeled by equating it to a series of juxtaposing rectangular prisms and calculating their thicknesses. By a mathematical tool named Taylor series non-linear problem changes to a linear problem near to initial model. Prepared algorithms here is baxsed on Levenberg-Marquardt method which by an iterative method comparing model response with actual data, will modify initial model. In this study, a sedimentary basin structure is equated to a series of juxtaposing rectangular prisms, whose horizontal dimensions are equal to the dimensions of the observation points grid and their density contrast is the constant density contrast between sediments and baxsement. The bottoms of the prisms coincide with the sediments and baxsement interface. There is one prism below each of the internal anomaly points. The inversion scheme calculates depths to bottom of each of these prisms iteratively using Levenberg-Marquardt method. Algorithm to invert gravity data : 1st, a function subprogram named “FORGRAV3D.m” calculates the gravity anomalies of initial model. 2nd, gravity anomalies of the initial model are calculated and compared with those observed. 3rd, a function subprogram named “JACOBIANGRAV3D.m” calculates the derivatives of gravity effects with respect to bottom depths of prisms. 4th, the difference between the observed and calculated anomalies are used to refine the thicknesses of the prisms. Modified model sets into step one as initial model to go into another iteration. Algorithm to invert magnetic data : 1st, a function subprogram named “FORMAG3D.m” calculates the magnetic anomalies of initial model. 2nd, magnetic anomalies of the initial model are calculated and compared with those observed. 3rd, a function subprogram named “JACOBIANMAG3D.m” calculates the derivatives of magnetic effects with respect to bottom depths of prisms. 4th, the difference between the observed and calculated anomalies are used to refine the thicknesses of the prisms. Modified model sets into step one as initial model to go into another iteration. The efficiency of the method and subprograms has been shown by inverse modeling of free noise and noise contaminated synthetic data both gravity and magnetic. Finally, we inverted gravity and magnetic field data from Abardezh area in central Iran that conclusion was acceptable.