Time-delay systems appear frequently in control engineering applications. Their presence should be considered both in analysis and synthesis of controllers. If they are not taken into account, the closed-loop system response may degrade until instability, particularly if the system to be controlled is unstable. Basically the control schemes that are used to stabilize time-delay systems can be classified into two groups: classical control approaches and control schemes specifically designed for time delay systems. This last approach is known in the literature as Dead-Time Compensators (DTC). With respect to that, it's worth mentioning the Smith Predictor (SP) and their modifications, and on the other hand the Finite Spectrum Assignment (FSA) technique. The main feature is that delay is eliminated from the characteristic equation of the closed-loop system. In this work new contributions for analysis and synthesis of controllers for discrete-time systems with input and output time-varying delays will be provided. Concretely, the goal consists of applying a control scheme based on the feedback of the future prediction of state (discrete implementation of FSA control scheme), known as predictor, by using the discrete-time process model, and comparing the achieved performance with respect to another control schemes proposed in literature. The good behavior of the predictor-based control scheme has been previously reported for time-constant delay systems. However, no contributions have been found for the time-varying case. Issues such as robust stability analysis against model and delay uncertainty, and the synthesis of the predictor-based control law will be discussed. Finally, with the aim to verify the achieved results, the predictor-based control law has been implemented in a real-time platform.