This paper investigates the actuator fault detection and tolerant control problem for vehicle platoons with partial loss of effectiveness and bias faults. Considering the inherent coupling between the control input and the nonlinear term in the actuator, a decoupled observer based fault detection algorithm with time-varying threshold is developed to decide whether the actuator fault occurs in the platoon. Then, by employing radial basis function neural networks (RBF NNs) and backstepping control method, an adaptive robust fault-tolerant control (FTC) scheme is proposed, which is proved to be capable of achieving the safety, stability and tracking performance of vehicle platoons in the presence of actuator fault. Simulation studies illustrate the effectiveness of the proposed method.