In order to identify the fault sensors and isolate them, a fault detection method for aero-engine servo sensor based on virtual redundancy was proposed. The Elman neural network trained by the modified algorithm was used to map the high-pressure speed to the main fuel in the whole envelope, and the virtual redundancy of main fuel was established. A servo sensor fault detection structure based on virtual redundancy was proposed, and a simulation of 2 failure modes which cannot be identified by conventional detection methods was used to verify the proposed method. The results show that the method can effectively identify the sensor failure, its structure is simple and the calculation is small.