Ammonia (NH3) poses significant risks to the environment and human health, serves as a precursor to nitrous oxide, and represents a source of nutrient loss and inefficiency. Quantification of NH3 emissions on farms is challenging because manure management practices and environmental conditions can vary widely. Process-based modeling offers a more robust approach than empirical methods for addressing this complexity. The Ruminant Farm Systems (RuFaS) model combines process-based models across the farm system to estimate whole-farm NH3 emissions. In particular, an established NH3 volatilization equation is used to assess emissions from the barn floor. This study develops a data assimilation algorithm to estimate farm-specific parameters for NH3 emissions. To enhance computational efficiency, we isolated the NH3 emissions method from the RuFaS model and conducted a first-order sensitivity analysis to identify key parameters for estimation. Among the parameters analyzed (manure pH, housing-specific constant, barn area, and manure density), manure pH had the highest sensitivity index (0.661), indicating its strong influence on NH3 emissions. Using this insight, we developed a Metropolis-Hastings algorithm to estimate manure pH and evaluated the method with simulated housing NH3 emissions data. The proposed data assimilation method consistently converged to the target pH value of 7.7, regardless of the initial values (6.5, 7.7, or 8.5), with an average acceptance rate of 21.5%. These results demonstrate the algorithm’s potential for accurately estimating farm-specific parameters. Future work will focus on extending the algorithm to estimate multiple parameters simultaneously and validating it with on-farm NH3 measurements. This work presents an opportunity to advance the development of regional NH3 emission inventories tailored to specific manure management practices.