Purpose <p>Exercise has multiple physical and mental health benefits for autistic youth. Dropouts from randomized controlled trials (RCTs) pose a threat to the validity of the evidence, with dropout rates varying across studies. We conducted a meta-analysis to investigate the prevalence and predictors of dropout rates among autistic youth participating in exercise RCTs.</p> Methods <p>Two authors searched major electronic databases up to 04/2025. We included RCTs of exercise interventions in autistic children and adolescents that reported dropout rates. A random effects meta-analysis, calculating the dropout prevalence in exercise group and a comparative meta-analysis comparing dropout in exercise versus control groups were conducted. Meta-regressions and subgroup analyses were conducted to explore dropout moderators.</p> Results <p>21 RCT’s involving 25 exercise interventions (<i>n</i> = 565, mean age 5.1 to 12.8 years; 81.2% males) were included. The trim-and-fill-adjusted prevalence of dropout across all studies was 24.4% (95%CI = 18.1% to 32.1%). Significant differences in dropout rates were found for the type of exercise intervention (<i>p</i> &lt; 0.001), intensity (<i>p</i> = 0.003), and overall intervention duration (<i>p</i> = 0.032). Dropout was similar in exercise versus control conditions (OR = 1.0, 95%CI = 0.97 to 1.0, <i>p</i> = 0.90, I² = 0%).</p> Conclusions <p>Exercise is well tolerated by autistic children and m-adolescents and drop out in RCT’s is comparable to control conditions. Thus, exercise is a feasible treatment and dropout rates decrease when the intervention is aerobic exercise, with either low or high intensity and when total intervention duration is shorter.</p>