Abstract: Objectives: Shipborne gravity anomaly data plays a crucial role in the construction of marine gravity field models. Although the existing shipborne gravity anomaly data have undergone conventional gravity measurement corrections and dynamic environmental effect corrections at the time of acquisition, the data span a long time period, involve multiple measurement agencies, gravity instruments, and reference frames, and thus exhibit a multi-source characteristic. Furthermore, due to the limitations of measurement and data processing techniques at the time of acquisition, the quality of the data varies and contains significant errors, making direct application difficult. Methods: In this study, a joint reprocessing method for multi-source shipborne gravity anomaly data is proposed, which refines the ship tracks through outlier elimination, long-wavelength error correction, intersection point adjustment, and systematic error compensation. First, rough tracks and remaining rough points were eliminated by comparing them with the reference gravity field; second, a quadratic polynomial model derived from the normal gravity formula was used to correct the long-wavelength error of the survey lines; finally, the error estimation and systematic error compensation of the observed values were performed through intersection point condition adjustment and the construction of a mixed polynomial model. Results: Taking the Gulf of Mexico as an example, shipborne gravity anomaly data in this region were obtained from the National Centers for Environmental Information (NCEI). After a series of joint processing methods, the root mean square (RMS) of crossover differences is reduced from 12.1mGal to 3.7mGal, and the residual RMS of gravity anomaly model SIO V32.1 is reduced from 6.62mGal before refinement to 3.91mGal. The difference between the two sets of ship tracks and the model was analyzed in the frequency domain, and the power spectral density of the error was significantly reduced in all frequency domains. Conclusions: As this study combined shipborne gravity data from different periods for joint processing, the final processing results have a certain gap compared to the precision of modern high-precision shipborne gravity measurements. However, overall, the accuracy of shipborne data has been significantly improved, which greatly enhances the utilization rate of shipborne data. This method can be used to further optimize global shipborne gravity anomaly data, providing a reliable dataset for constructing high-precision ocean gravity field models.