This paper presents an attracting law approach based on the adaptive switching gain for discrete-time systems with input/output description. This method can automatically tune the switching gain according to the vary speed of closed-loop system uncertainties, and directly presents properties of the error-dynamics. In order to characterize both the convergence performance and steady-state performance, the steps for tracking error to converge to the origin, and the expressions for the range of the steady-state error and the absolute attractive layer are derived. Both numerical simulation and experiment results show the effectiveness of the proposed control method.