The performance of deadbeat predictive current control algorithm depends on the accuracy of the motor parameters used to build the prediction model; however, due to the influence of temperature variation, magnetic saturation, cross-coupling, internal unmodeled and external unknown disturbances, motor parameter uncertainty inevitably occurs during motor operation, which affects the steady-state static error and dynamic fast response capability of the DPCC algorithm. In addition, the control delay of the digital control system, including current sampling delay, duty cycle refreshment delay and other factors, will also greatly reduce the control performance of the system. Therefore, in order to solve the appeal problem, this paper proposes a dual-observer current prediction control method without differential beat based on Luenberger state observer and sliding mode perturbation observer, which mainly estimates the perturbations caused by parameter changes and feed-forward compensates the control voltage through the dual-observer, and at the same time, utilizes the predicted current at the next moment to replace the actual current value at the current moment in the predictive control model to compensate the control delay, and at the same time, resolves the parameter uncertainty and the duty-cycle refresh delay. At the same time to solve the parameter uncertainty and the disturbance caused by the control delay. Simulation experiments prove that this method can have good parameter uncertainty suppression effect, improve the dynamic current following performance and reduce the steady state error, and the current ripple is also significantly reduced, improving the control performance of the motor.