Projections of sea level change require ice sheet models that reconcile observations with ice flow dynamics to constrain poorly observed processes and parameters. We use the Ice-sheet and Sea-level System Model (ISSM) to construct a state and parameter estimate of Ronne Ice Shelf (RIS) and its drainage basin from 1992 to 2018. We assimilate time-varying observations of surface velocity and elevation change to reconstruct a dynamically consistent model of time-mean basal friction and ice rheology, and sub-shelf basal melting rates. Relative to existing methods, we reduce uncertainty weighted root mean square errors (RMSE) in elevation change and velocity by 21.2 % and 43.6 %, respectively. Our inferred basal melt estimates capture spatio-temporal patterns driven by ocean currents at 5 km spatial resolution, and match monthly amplitudes 43.1 % closer to observations relative to existing estimates. Detrended mass change in the upstream grounded basin also correlates with Gravity Recovery and Climate Experiment (GRACE-GFZ) estimates (<em>R</em> = 0.754). This physically-consistent 26-year reconstruction provides a framework for applications to Antarctica and beyond, with implications for prognostic modeling and future sea level projections.