Abstract: Based on the full three dimensional optimization software of NUMECA/Design 3D, both artificial neural network (ANN) and genetic algorithm (GA) were used to optimize axial flow rotor by using the three dimensional combination of lean and sweep stacking line. The optimization objective is to maximize efficiency and total pressure rise at design operation point while maintaining mass flow. Results revealed that, based on the three dimensional combination of lean and sweep optimization, the overall aerodynamics performance of the optimized rotor was improved. It is found that the efficiency of the optimized rotor is increased by 1.92% and its total pressure is increased by 3.98% at design operation point. The composited optimization of lean and sweep could adjust load distributions on the blade surface, improve the flow near the blade tip and hub, and suppress the flow separation near the corner area of blade root. Furthermore, the optimization reduced the strength and the influence range of the tip leakage vortex, which reduced the flow loss and enhanced the blade working ability.