This paper presents an experimental comparison of the Kalman and complementary filter for attitude estimation of a microwave quadrotor, fusing measurements from inertial MEMs devices. Simulations on Kalman and complementary filters are performed to study how the parameters affect the estimations of each one. Finally, the experimental results are presented using real data, which are used for both noise parameter modelling and comparing the performance of the filters given different operating conditions of the experimental platform.