Abstract: To obtain the variation of electrochemical performance of direct ammonia solid oxide fuel cell (DASOFC) with operating conditions, a three-dimensional DASOFC simulation model was established. An artificial neural network (ANN) was used for performance prediction, and the particle swarm optimization algorithm and NSGA-II algorithm were employed to obtain the optimal operating parameters of DASOFC under different conditions. The results show that under different power densities, the difference between the maximum and mean fuel utilization ranges from 4.10% to 53.08%. Under different fuel utilizations, the difference between the maximum and mean power density ranges from 0.139 9 to 0.559 0 W/cm². Under different operating voltages, the optimization effects on fuel utilization and power density are not significant. Under different inlet flow rate conditions, the differences between the optimal and mean values of fuel utilization and power density are 24.81%-51.93% and 0.328 0-0.844 7 W/cm², respectively. Under different inlet temperature conditions, the differences between the optimal and mean values of fuel utilization and power density are 18.11%-30.89% and 0.247 5-0.682 8 W/cm², respectively. Adjusting the operating voltage as the primary approach, supplemented by controlling the anode inlet flow rate and inlet temperature, is an effective means to improve power density and fuel utilization.