BackgroundExercise has shown promise for supporting executive function in children with autism spectrum disorder (ASD), but findings remain inconsistent, partly because exercise dose has not been examined systematically. This study examined the immediate effects of acute cycling with different combinations of intensity and duration on executive function in children with ASD.MethodsA within-subject repeated-measures design was used. Thirty children with ASD completed one non-exercise sedentary control session at enrollment, followed by four counterbalanced cycling sessions arranged in a 2 × 2 exercise-dose matrix: 20 min low-intensity, 40 min low-intensity, 20 min moderate-intensity, and 40 min moderate-intensity cycling. Exercise intensity was monitored using the heart rate reserve method. Selected executive function components were assessed after each condition using the Early Years Toolbox, including the Go/No-Go task for inhibitory control and the Mr. Ant task for working memory. Repeated-measures analyses of variance were conducted, followed by FDR-corrected planned comparisons between each exercise condition and the sedentary control condition.ResultsSignificant condition effects were observed for Go reaction time (p = 0.024, partial η2 = 0.092), No-Go reaction time (p = 0.006, partial η2 = 0.116), Go accuracy (p = 0.046, partial η2 = 0.080), No-Go accuracy (p = 0.006, partial η2 = 0.115), and the inhibitory-control composite score (p < 0.001, partial η2 = 0.405). Moderate-intensity exercise showed the clearest post-session advantages relative to sedentary control, with the 40-min moderate-intensity condition showing the most consistent pattern across inhibitory-control outcomes. For the inhibitory-control composite score, all four exercise conditions were higher than sedentary control. No significant condition effect was found for working memory (p = 0.353, partial η2 = 0.037).ConclusionAcute cycling may have selective short-term effects on selected executive function components in children with ASD. Moderate-intensity cycling, particularly a 40-min session under the present protocol, was associated with more favorable inhibitory-control performance, whereas immediate changes in working memory were not observed. These findings should be interpreted in light of the non-counterbalanced sedentary control session.Clinical trial registrationhttps://www.chictr.org.cn/showproj.html?proj=248491, ChiCTR2400091872.