In this paper, a new method based on a genetic algorithm and Minkowski Island fractal is proposed for multiband antennas. Three-antenna configurations are chosen to validate the proposed optimization procedure. The first configuration is a wide-band antenna, operating in the WLAN (wireless local area network) UNII-2C band. The second configuration is a dual-band antenna, operating in the WLAN UNII-2 and UNII-2C bands. In contrast, the third is a tri-band antenna operating in the UNII-2, UNII-2C, and UNII-3 bands. The optimization process is accelerated by using the Computer Simulation Technology (CST) Application Programming Interface which allows all genetic operators to be performed in MATLAB while the numerical calculations are running in the internal CST Finite-Difference Time-Domain -solver using parallel computing with GPU acceleration. All three designed configurations are manufactured using a $\textstyle0.8\;\text{mm}$ thick FR4 epoxy substrate with a relative dielectric constant of $4.8$. The return loss and the radiation pattern’s measurements agree well with the simulation results. Further, the methodology presented can be very effective in terms of size reduction; the designed antennas are $24 \times 24 \times 0.8\;{\textrm{m}}{{\textrm{m}}^3}$ ($460\;{\textrm{m}}{{\textrm{m}}^3}$).