Reducing glycaemic index (GI) and glycaemic load (GL) inconsistently improves aspects of cognitive function and appetite in children. Whether altering the GL by lowering carbohydrate relative to protein and fat has a role in these effects is unknown. Therefore, we assessed the differential effects of beverages varying in GL and dairy composition on appetite, energy intake and cognitive function in children. A total of forty children (10–12 years) completed a double-blind, randomised, crossover trial, receiving three isoenergetic drinks (approximately 1100 kJ): a glucose beverage (GI 100, GL 65), a full milk beverage (GI 27, GL 5) and a half milk/glucose beverage (GI 84, GL 35). For 3 h post-consumption, subjective appetite and cognitive performance (speed of processing, memory, attention and perceptual speed) were measured hourly. At completion, each child was provided a buffet-style lunch and energy intake was calculated. Blood glucose was objectively measured using the Continuous Glucose Monitoring System. Blood glucose AUC values were significantly different between the drinks (P< 0·001), but did not sustain above the baseline for 3 h for any drink. Mixed modelling revealed no effect of beverage on subjective appetite or energy intake. Participant sex and drink GL significantly interacted for short-term memory (P< 0·001). When girls consumed either milk-containing beverage, they recalled 0·7–0·8 more words compared with 0·5 less words after the glucose drink (P≤ 0·014). Altering GL of drinks by reducing carbohydrate and increasing protein did not affect appetite or cognition in children. Girls may demonstrate improved short-term memory after consuming beverages with higher protein and lower GL.