In this paper, an adaptive decentralized PID design is developed for multivariable systems. In the proposed design, the controller parameters are adjusted by an updating algorithm derived based on the Lyapunov method such that the predicted tracking error converges asymptotically. Toward this end, the Just-in-Time Learning method is incorporated into the proposed design to provide the information required for the updating algorithm. Simulation results illustrate that the proposed design achieves better control performance than its corresponding benchmark designs reported in the literature. (C) 2012 Elsevier Ltd. All rights reserved.