For the system with model uncertainty, an active fault detection method based on auxiliary signals and redundant actuators is proposed to detect micro faults. Firstly, the optimal observer is constructed to obtain the residual signal by improving the fault sensitivity and anti-interference robustness. Then, an auxiliary input signal is introduced to increase the residual error of the fault system. At the same time, a redundant actuator is constructed to avoid the influence of the auxiliary input signal on the fault free system. In addition, an adaptive threshold construction method is proposed to reduce the conservatism of fault detection. When the fault occurs, the residual changes rapidly and is higher than the threshold value to achieve fast detection. Finally, the simulation results verify the effectiveness of the proposed method.