This paper presents an optimization method for fan/compressor which couples throughflow model solving axisymmetric Euler equations with adaptive simulated annealing (ASA) algorithm. One of the advantages of this optimization method is that it spends much less time than 3D optimization due to the rapid solving of throughflow model. In addition, the optimization space is quite extensive because more design variables can be adjusted in throughflow phase, such as swirl distribution, hub curve and sweep. To validate this optimization method, a highly loaded fan rotor with pressure ratio of 3.06 as a baseline is optimized. During the optimization process, the objective function is total pressure ratio, moreover, mass flow and efficiency are selected as the constraint conditions. Three important design variables including swirl distribution, hub curve and sweep are parameterized using Bezier curve, and then optimized in throughflow model independently, finally the optimum designs are validated using 3D viscous CFD solver. It is shown that pressure ratio and rotor loading can be improved further through optimizing swirl distribution, however, hub and sweep curves take more effects on mass flow and efficiency respectively. The optimization results demonstrate the advantage and feasibility of this optimization method.