In this study, we examined the potential of various artificial bee colony (ABC) algorithm variants for optimizing the design of fractional-order Proportional Integral and Derivative (FOPI and FOPID) controllers for multi-input-multi-output systems. The performance of the FOPI and FOPID controllers was evaluated in controlling Wood and Berry distillation columns and an industrial polymerization reactor with and without decoupling. The simulation results showed that the multi-population ABC algorithm based on global and local optima outperformed the other ABC algorithm variants and other state-of-the-art optimization algorithms in terms of control performance.