Due to the small size and weight of micro aerial vehicle (MAV), only miniaturized MEMS type sensors are applicable for MAV autonomous flight system. In this paper, we propose a accelerometer and gyro mixing algorithm to improve an attitude performance of MEMS type sensors. The performance of the proposed mixing algorithm is compared with the performance of fuzzy-based mixing algorithm through simulation. The simulation results show that the attitude compensation method through the attitude compensation has better performance than the fuzzy-based mixing method for MAV attitude estimation.