Time-growing navigation error due to inevitable measurement errors in the MEMS-grade inertial sensor is one of the main challenges in low-cost inertial navigation systems (INSs). This paper aims to develop AHRS/GPS/DR integrated navigation algorithm for long-range autonomous underwater vehicle (AUV). Proper performance in deals with long-term GPS outage is the main advantage of the proposed low-cost navigation algorithm. This is due to appropriate estimation of gyro biases in the proposed algorithm. Using Dead Reckoning (DR) method based on an axial speed sensor, an applied algorithm is developed for position estimation during the GPS outages. The algorithm is evaluated by field test executed on a long-range AUV with high underwater durability. Experimental results indicate that in spite of 500 seconds GPS outage, the estimation error in the heading angle does not exceed 5 degrees.