<p>Selective Catalytic Reduction (SCR) is a key technology for controlling nitrogen oxide (NO_x) emissions from diesel engines. In this study, the effects of urea injection parameters on SCR system performance are systematically investigated using computational fluid dynamics (CFD). Grid independence and model validation are first conducted to ensure numerical reliability. Single-factor analyses are then performed to examine the influences of injection speed, droplet diameter, and dispersion coefficient on SCR performance. The results show that injection speed and droplet diameter mainly affect droplet evaporation and wall impingement behavior, whereas the dispersion coefficient plays a dominant role in spray distribution and mixing uniformity. Based on these results, response surface methodology (RSM) is employed to establish regression models and perform local optimization. In addition, the Non-dominated Sorting Genetic Algorithm II (NSGA-II) algorithm is applied for global multi-objective optimization, and the optimal solution is determined from the Pareto front using the Analytic Hierarchy Process-Technique for Order Preference by Similarity to Ideal Solution (AHP-TOPSIS method). A comparative analysis indicates that RSM exhibits higher predictive accuracy for de-NO_x efficiency, while the NSGA-II-based approach shows superior performance in reducing urea deposition. These findings provide practical guidance for the optimization of urea injection parameters and deposition control in SCR systems.</p>