In the light of worse control caused by sensor faults in the process of adjusting the ride height for electronically controlled air suspension (ECAS) system, a method that can diagnose the sensor faults for ECAS system is presented. A physical model of the ECAS system is built by AMESim in order to describe accurately the air spring characteristics. Meanwhile，mathematical models of road excitation and sensor faults are constructed by Matlab/Simulink. For the nonlinear characteristics of the ECAS system, extended Kalman filters, which are designed by extended Kalman estimation algorithm, are used to establish the residual observer bank of the sensor fault. According to the state estimations achieved by extended Kalman filter bank, output residuals are obtained and could be compared with a threshold to detect and isolate the sensor faults. Then, the co-simulation of different sensors with different faults is carried out. Finally, the test bench of 1/4 ECAS system is built to perform experiments of sensor fault diagnosis during the process of height adjustment. Simulation and test results show that the proposed approach accurately detect typical sensor faults of the ECAS system and preferably isolated different sensors with faults to ensure the accurate and reliable operation of the ECAS system.