In this work, we apply classical and new self-adaptive differential evolution (DE) algorithms to the design of optimum gain pyramidal horns. The application of DE algorithms to pyramidal horn design allows the accurate calculation of their physical dimensions in a short amount of time. Moreover, self-adaptive DE does not require the prespecified choice of control parameters reducing significantly the users' effort. Representative examples demonstrate the applicability of our proposal. Our study shows that self-adaptive DE algorithms outperform or produce similar results with other methods in the literature in terms of solution accuracy and convergence speed.