A new optimization scheme combined with the space-mapping technology and the jumping-gene genetic algorithm (JGGA) is to replace the conventional direct genetic computational type of the optimization scheme for antenna designs. This utilizes the assumption that there is an equivalent linear mapping between the design-parameter ratio and the response ratio in two spaces on the condition that there is only a difference in simulation mesh setup between coarse and fine models of antennas. The design parameters and responses of an optimal individual in coarse space at each generation are collected during the JGGA parameter extraction. The above linear mapping is then used to produce candidates of new design parameters in fine space. A multicriteria decision-making selection strategy is employed to govern the optimization process. This scheme is applied to a novel radio frequency identification tag antenna with an artificial magnetic conductor ground. The operation results of the combination algorithm show great improvement in the optimization when compared with the previous work in the same field. Furthermore, the proposed antenna can show relatively tolerant reading distances in various platform environments through experiments.