With the optimization design, the non-axisymmetric end wall in a high pressure(HP) turbine stator row was optimized based on the methods which combined end wall parameterization, three-dimensional Navier-Stokes(N-S) flow computation, and genetic algorithm(GA) ground on artificial neural network(ANN). The objective of optimization design is to minimize the total pressure loss coefficient and secondary kinetic energy(SKE) at the stator exit while the inlet mass flow, the exit Mach number and the exit flow angle are controlled. This paper compared and analyzed the influences on the parameters at the stator exit and the stage performance of the HP turbine between the axisymmetric end wall before optimization and the non-axisymmetric end wall after optimization. The analysis results indicate that the optimized non-axisymmetric end wall can effectively improve the flow field through suppressing the development of secondary flow vortex system in the HP turbine stator so as to reduce the flow loss at the stator exit. And the total pressure loss coefficient at the stator exit reduces by 14.85%, while the stage isentropic efficiency increases by 0.456%.