A systematic review and Bayesian network meta-analysis is necessary to evaluate the efficacy and safety of triple therapy with different doses of inhaled corticosteroids (ICS) in stable chronic obstructive pulmonary disease (COPD). We selected 26 parallel randomized controlled trials (41,366 patients) comparing triple therapy with ICS/long-acting beta-agonist (LABA), LABA/long-acting muscarinic antagonist (LAMA), and LAMA in patients with stable COPD for ≥ 12 weeks from PubMed, EMBASE, the Cochrane Library, and clinical trial registries (search from inception to June 30, 2022). Triple therapy with high dose (HD)-ICS exhibited a lower risk of total exacerbation in pre-specified subgroups treated for ≥ 48 weeks than that with low dose (LD)-ICS (odds ratio [OR] = 0.66, 95% credible interval [CrI] = 0.52–0.94, low certainty of evidence) or medium dose (MD)-ICS (OR = 0.66, 95% CrI = 0.51–0.94, low certainty of evidence). Triple therapy with HD-ICS exhibited a lower risk of moderate-to-severe exacerbation in pre-specified subgroups with forced expiratory volume in 1 s < 65% (OR = 0.6, 95% CrI = 0.37–0.98, low certainty of evidence) or previous exacerbation history (OR = 0.6, 95% CrI = 0.36–0.999, very low certainty of evidence) than triple therapy with MD-ICS. Triple therapy with HD-ICS may reduce acute exacerbation in patients with COPD treated with other drug classes including triple therapy with LD- or MD-ICS or dual therapies.