To effectively control and maintain the transient stability of power systems, traditionally, the extended Kalman filter (EKF) is used as the real-time state estimator (RTSE) to provide the unmeasurable state information. However, the EKF estimation may degrade or even become unstable when the measurement data are inaccurate through random sensor failures, which is a widespread problem in data-intensive power system control applications. To address this issue, this paper proposes an improved EKF that is resilient against sensor failures. This work focuses on the resilient EKF’s (REKF’s) derivation with its application to single-machine infinite-bus (SMIB) power system excitation control. The sensor failure rate is modeled as a binomial distribution with a known mean value. The performance of REKF is compared with the traditional EKF for power system observer-based control under various chances of sensor failures. Computer simulation studies have shown the efficacy and superior performance of the proposed approach in power system control applications.