In this article, the fault-attack control problem is investigated for wheeled mobile robot (WMR) systems subjected to actuator faults, disturbances, communication attacks, and limited communication resources. An event-observer based compensation controller is presented. With the help of the observer estimation values and the attack sleep/active instant trigger information, the tracking control performance is realized for the robot system with the assistance of the neural network approximation technology. Concretely, first, the robot system dynamic model with actuator faults, disturbances, and attacks is established. Then, an event-based proportional–integral observer (PIO) is established. In the observer framework, a state estimator, an actuator fault efficiency estimator, and a disturbance estimator are embedded. Based on the observer outputs, a second-order adaptive sliding mode fault-compensation reliable controller is presented. In this controller framework, the fault compensation, disturbance attenuation, and the attack sleep/active time instant information are contained to guarantee the reliability and performance recoverability of the robot system. Furthermore, a dynamic even condition and an adaptive trigger scheme are constructed in the sensor and actuator channel to achieve the communication-efficient purpose. Finally, two cases of the robot system are performed to verify the system recoverability of the presented approach.