The selective non-catalytic reduction(SNCR) performance is sensitive to the process parameters such as flow velocity, reaction temperature and mixing of reagent(ammonia or urea) with the flue gases. Therefore, the knowledge of the velocity field, temperature field and species concentration distribution is crucial for the design and operation of an effective SNCR injection system. In this work, a full-scale two-dimensional computational fluid dynamics(CFD)-based reacting model involving a droplet model is built and validated with the data obtained from a pilot-scale urea-based SNCR reactor installed with a 150 kW LPG burner. The kinetic mechanism with seven reactions for nitrogen oxides(NOx) reduction by urea-water solution is used to predict NOx reduction and ammonia slip. Using the turbulent reacting flow CFD model involving the discrete droplet phase, the CFD simulation results show maximum 20% difference from the experimental data for NO reduction. For NH3 slip, the simulation results have a similar tendency with the experimental data with regard to the temperature and the normalized stoichiometric ratio(NSR).