A simple and efficient method was developed to improve soil moisture representation in an operational water balance model through satellite data assimilation. The proposed method exploits temporal covariance statistics between modelled and satellite-derived soil moisture to produce analysed estimates, as a weighted combination of all data sources. We demonstrate the application of the method to the Australian Water Resources Assessment (AWRA) model and evaluate the accuracy of the approach against in-situ observations across the water balance. The correlation between simulated surface soil moisture and in-situ observation is increased from 0.54 (open-loop) to 0.77 (data assimilation). We suggest an approach to use analysed surface moisture estimates to impart mass conservation constraints on related states and fluxes of the AWRA model in a post-analysis adjustment. The improvements gained from data assimilation can persist for more than one week in surface soil moisture estimates and one month in root-zone soil moisture estimates.