The boost converter is widely used in voltage regulation applications and active power factor correction (PFC), its proper operation and performance are directly related to final design of its control loop. The design of this control is somewhat complex due to the nonlinearity of its behavior; the commutation of a switch (power transistor) makes the equivalent circuit change continually over time (two alternating circuits in time). In this work, it is defined the poles of the transfer function in closed loop for the average model (linear model approximate) of the boost converter using evolutionary principles of genetic algorithms, with the idea of improving the dynamic response of voltage regulator. The design outline is successfully verified by simulation and on a laboratory prototype, and improvements were found in the dynamic (voltage overshoot and response time) compared to classical schemes.