Abstract: For processing the data from the gravity survey of the North sea,Strang van Hees(1983) once proposed to use the collocation method for the adjustment of the cross-points.Prince and Forsyth(1984) presented a simple objective method for minimizing crossover errors in marine gravity data.They suggested that the track-lines be divided into straight-line segments and a least squares approach be used to calculate the individual segment corrections.On the basis of an overall error analysis and assessment on the Lamont-Doherty Geological Observatory's global marine gravity data bank,Wessel and Watts(1988) introduced a method to reduce the standard deviation of crossover discrepancies.In order to produce a new free-air gravity anomaly map around Japan,Adjaout and Sarrailh(1997) collected and processed the marine gravity measurements made by ships separated in time by a number of years over the area.In data processing,they also used a crossover minimization method to calculate the linear drift and the biases.The adjustment was performed by constraining the biases to a zero sum and/or by including reference station constraints.According to an analysis of physical and geometric uncertainties which caused crossover errors,the first authour of this paper put forward a new method of combined adjustment which took into account both direct and indirect influences of ship navigation in 1995.In 1999,Huang et al.discussed again the crossover adjustment in a more extensive senseand provided a method of self-calibrating surveying line network adjustment.The purpose of this paper is to simplify the rigorous methods mentioned above for the consideration of practical application.A method of two steps processing is put forward in this paper.