The actuator reliability is of great significance to the system security. Therefore, a fault isolation and identification method is proposed for autonomous underwater vehicles(AUVs) based on fault factor and multiple observers scheme. Firstly, the fault isolation and identification of AUV actuators is divided into two levels based on the concept of fault factors. Then, the fault isolations focusing on rudders, propellers and actuators based on extended state observers(ESOs) and isolation logic rules are proposed. According to the relationship between the output force and torque, further identification can be realized via internal information analysis of fault factors. The simulation results show that the proposed scheme has an outstanding accuracy for AUV actuator fault isolation and identification.