Abstract: For shortcomings of control methods of offline and online data-driven, the paper presented a control method of offline-online hybrid data-driven. First it gave an offline data-driven control method based on the least squares support vector machine and virtual reference feedback tuning, then gave control structure and design steps of the control method of offline-online hybrid data-driven combined with model-free adaptive control in the framework of two degrees of freedom control system. The proposed method skips the process of charged object modeling, greatly reduces cost of controller design, and avoids introduction of model error. Applying the method in offline-online data-driven of DC motor, the simulation result showed that the method can achieve target signal tracking and uncertainty (external disturbance) inhibiting of motor system effectively.