A robust multiple-model adaptive control is applied for a launch vehicle, consisting of mixed-µ regulators, each designed for a certain region of uncertainty, and an identification subsystem. The identification subsystem is composed of a bank of Kalman filters and an algorithm that calculates the posterior probabilities of uncertain process model parameters, switching or mixing the outputs of the controllers. The algorithm is tested for different variations of the flight parameters during the trajectory of the atmospheric phase, validating the control system in terms of performance, stability and robustness.