This study was conducted over 2 years (2011 and 2012) to determine the optimal combinations between the irrigation rate and glycinebetaine (GB) levels in order to maximise yield and irrigation water use efficiency (IWUE) for drip-irrigated maize. A field experiment was performed using a randomised complete block split plot design with three drip irrigation rates (I-1: 1.00, I-2: 0.80, and I-3: 0.60 of the estimated evapotranspiration, ET) and five GB levels (GB(0), GB(25), GB(50), GB(75) and GB(100), GB levels at 0, 25, 50, 75 and 100 mM, respectively) as the main and split plots, respectively. We found that although exogenously applied GB appeared to have different effects on yield variables and IWUE, these differences were dependent on the level of GB within the same irrigation rate. The grain yield and yield component values for I(2)GB(50) treatment were occasionally comparable to those obtained for I(1)GB(0) treatment, and the values for both treatments were higher than those obtained for I(1)GB(75) or I(1)GB(100). I(3)GB(50) or I(3)GB(100) had grain yield and yield component values similar to those obtained for I(2)GB(0) and I(2)GB(100). The highest value for IWUE was found for I(2)GB(50) and this value was similar to that obtained with I(3)GB(75), while the lowest values were obtained for I(1)GB(75) or I(1)GB(100). Medium GB levels were effective under I-2 and I-3 treatments to obtain the lowest value for seasonal yield response factors (k(y)). The production functions of yield versus GB levels were second-order relationship for all drip irrigation rates. In conclusion, exogenous application of GB has the potential to improve yield and IWUE under limited water application, while a threshold level of GB was required for a positive effect. (C) 2014 Elsevier B.V. All rights reserved.