Persistently elevated glycosylated hemoglobin (HbA1c)is associated with a higher risk of long-term vascular complications. We evaluated the effect of different exercise modalities and doses on HbA1c levels in patients withtype 2 diabetes. A systematic search for randomized controlled trials involving exercise interventions in patients with type 2 diabetes was conducted across seven electronic databases, encompassing data from their inception up to October 2023. Two independent reviewers assessed the quality of the literature. Pairwise, network, and dose-response meta-analyses using the random-effects model were conducted to analyze the effect of exercise on HbA1c levels. A total of 85 randomized controlled trials with 5535 participants were included. The network meta-analysis showed that high-intensity interval training(HIIT) has the highest ranked (MD = - 0.78% [- 8.50 mmol/mol]; 95% CrI: - 1.04, - 0.51), followed by combined aerobic and resistance exercise(CE) (MD = - 0.54% [- 5.90 mmol/mol]; 95% CrI: - 0.69, - 0.40), yoga (MD = - 0.45% [- 4.92 mmol/mol]; 95% CrI: - 0.77, - 0.13), and continuous aerobic exercise(CAE) (MD = - 0.42% [- 4.58 mmol/mol]; 95% CrI: - 0.54, - 0.30). In addition, a significant improvement in HbA1c levels can be observed when the totalexercise dose reaches 840 metabolic equivalent of tasks-min/week. There was low-quality evidence that HIIT, CE, yoga, and CAE may be effective treatment modalities for improving HbA1c in patients with type 2 diabetes, and there was no significant difference in efficacy. Moreover, a non-linear dose-response relationship was found for total exercise and HbA1c levels.Future research should further investigate the specific effects of different exercise doses on HbA1c levels in patients with type 2 diabetes and provide a more personalized exercise prescription.