With the widespread application of high-power motors in energy systems, axial fans have become a key method for thermal management. However, as motor power increases, the resulting temperature rise renders the traditional approach of enlarging the fan impeller diameter increasingly ineffective and may even degrade cooling performance. This study takes a 5800 kW nuclear power main feedwater pump motor as the research object, constructing a three-dimensional fluid field model of the axial fan end to analyze how structural features affect airflow direction and velocity distribution. A forward and reverse analysis method is proposed and applied to identify the influence of structural parameter variations on airflow velocity and temperature rise. This method enables the extraction of key structural factors in complex systems where clear functional relationships are lacking, thereby enhancing optimization efficiency. Parameter optimization improved the favorable flow velocity ratio and reduced temperature rise. Finally, the combined use of temperature field simulations and prototype testing verified the effectiveness of both the model and the proposed method. The findings provide a theoretical foundation and engineering reference for optimizing the airflow structure in high-power axial fan motors.