Abstract The ongoing melting of the Greenland ice sheet (GrIS) is a major contributor to global sea-level rise. Accurately quantifying this contribution requires reliable estimates of surface mass balance (SMB), yet substantial discrepancies persist among existing estimates due to observational limitations and model uncertainties. This study presents a particle batch smoother (PBS) data assimilation framework designed to improve SMB estimation by integrating model output with surface elevation change observations. The method was evaluated within an observing system simulation experiment using synthetic elevation changes emulating the NASA’s Ice, Cloud, and land Elevation Satellite-2 (ICESat-2) surface elevation change product. The results show that assimilating elevation change data significantly reduces errors associated with model physics and meteorological forcing. Compared to a nature run (assumed truth), the PBS framework reduces SMB bias by 65% and root-mean-square error by 84%. These findings demonstrate that the PBS approach can substantially enhance SMB estimates across the GrIS, providing a promising pathway toward improved reanalysis and projections of ice-sheet mass balance.