This paper presents an optimization technique that aims at multiband antenna design. The proposed method is based on the framework of multi‐objective evolutionary algorithms, and a two‐stage mechanism that balances the degree of optimizing impedance matching and the degree of providing a wide impedance bandwidth is incorporated. Conventionally, the design optimization of multiband antennas relies on minimizing the maximum reflection coefficient or maximizing the area under the return‐loss curve over targeted frequency bands. However, these widely used methods direct an optimization algorithm to improper solution sub‐domains in the multiband design problem. To overcome the limitation of these conventional methods, the general rule of objective functions is thoroughly investigated in this paper. Furthermore, a two‐stage optimizer is designed based on what the multiband optimization problem needs. With the use of the proposed method, two multiband antennas for mobile communication systems covering 824–960 MHz and 1710–2170 MHz are successfully developed. Simulated and measured results show that the proposed technique outperforms conventional optimization approaches significantly.