The problem of online adaptive tuning of PI controllers in heating, ventilating and air conditioning (HVAC) systems is explored. A discharge air temperature (DAT) system is described by a first-order plus dead-time model. The model parameters were estimated online while the control remains in the closed-loop. H-infinity adaptive tuning rules were applied and new PI parameters were estimated. Both computer simulations and experimental responses were studied. The results show that the H-infinity adaptive PI (HIAPI) controller is able to track set-point changes, reject the effects of multiple disturbances and is robust to changes in model parameters. (C) 2010 Elsevier Ltd. All rights reserved.