An automatic gravity inversion technique in the space domain simultaneously estimates the parameters of strike-limited listric fault sources and regional gravity background from a set of observed Bouguer gravity anomalies. The fault profile and regional effect are described by unknown polynomial functions of arbitrary but prescribed degree. Furthermore, the density contrast within the fault structure is presumed to be known, according to a prescribed parameterized nonlinear function of depth, in geologic settings where the detached downthrown block consists of a series of sedimentary beds whose density increases with depth. The inversion is automatic in that it initializes and determines polynomial coefficients for the fault boundary and regional gravity background from a set of observed Bouguer gravity anomalies and improves them iteratively until the modeled gravity anomalies mimic the observed anomalies. An analysis of a set of gravity anomalies attributable to a synthetic model of a listric fault structure in the presence of pseudorandom noise with and without regional background has disclosed that the algorithm yields reliable interpretations with modest error in model geometry, even in the presence of pseudorandom noise. In the presence of regional gravity background and pseudorandom noise, the estimated parameters of the structure deviate marginally from the true ones. The derived density-depth model of the Jharia coal basin in India, a pull-apart basin, has been used to analyze the observed Bouguer gravity anomalies of a boundary fault. The interpretation has yielded information consistent with drilling results and geologic setting of the basin.