<p>In this paper, an adaptive fault-tolerant learning control (AFTLC) method with data dropout prediction (DDP) is proposed for unknown nonlinear networked systems under nonrepetitive Denial-of-Service (DoS) attacks, actuator faults, and unknown control directions. In the considered system, the DoS attacks are carried out by intentionally blocking the transmission of the states between sensors and controllers and the transmission of control signals between controllers and actuators as well. Thus, the nonrepetitive of DoS attacks will result in nonrepeatable data loss of system states and control signals, severely degrading the tracking performance. A novel DDP mechanism, which updates based on the iterative error constructed by the predicted values, is designed to mitigate the impact of DoS attacks effectively. Then, a fuzzy logic system (FLS) is employed to approximate the desired control signal in the presence of DoS attacks, actuator faults and unknown control directions. After that, integrating with the DDP mechanism, FLS and discrete Nussbaum-type function technique, an AFTLC method is proposed in this paper, which can dynamically detect the correct control direction and compensate for the side effects caused by nonrepetitive DoS attacks and actuator faults. Finally, the convergence of the proposed AFTLC method is proved via composite energy function and simulation examples are provided to demonstrate its effectiveness.</p>