Abstract This paper presents the development of two dual-loop control strategies for an underwater Remotely Operated Vehicle, ROV. The first strategy consists of two PID controllers while the second one features a LQG controller in the inner loop, and a PID controller in the outer loop. The transient response of the two strategies is compared when an anti-windup gain is applied and when there is no anti-windup gain, and a way to tune this gain in the case of statespace controllers is proposed. For further comparison, the strategies are simulated for a variable set point, for the three variables to be controlled, x, z, yaw , so that both their tracking and the effect of coupling can be seen. In conclusion, the use of anti-windup gain is recommended for control systems that reach saturation, so that they can react quickly when they are in an operating zone, and the use of a random matrix for the calculation of this gain in state-space controllers is suggested.