To effectively mitigate NOX emissions in pulverized coal combustion, Computational Fluid Dynamics (CFD) is frequently employed as a predictive tool. Because fuel-bound nitrogen is the primary source of NOX in these applications, modelling largely relies on coal characterization—data that are often uncertain even with costly experimental measurements.This study introduces a systematic framework for estimating missing fuel‑nitrogen characteristics from standard proximate and ultimate analyses based on correlations from rapid coal devolatilization experiments. Utilizing a conventional CFD approach, we simulated coal combustion in a 150-kW facility, which mimics rotary kiln conditions. A sensitivity analysis evaluates the impact of uncertainties in volatile nitrogen yields, precursor compositions (HCN/NH3), and char surface reactivity on NOX predictions. Results indicate that input uncertainties may account for up to 25% of the total predicted NOX. The ratio of nitrogen released as volatiles versus char-bound nitrogen and the initial char surface area of the coal particles are both critical parameters, while variations in volatile precursor ratios had a negligible effect (<1%). The model was validated against measured NOX emissions from the 150-kW facility, demonstrating that while the framework provides a robust estimation tool, high-fidelity characterization of char properties remains vital for precise emission modelling.