Background <p>Pulmonary arterial hypertension (PAH) and related pulmonary hypertension populations are associated with impaired exercise capacity and reduced quality of life. Supervised exercise rehabilitation is recommended for clinically stable patients, but the relative effects and certainty of evidence for different exercise components and combinations remain uncertain.</p> Methods <p>We conducted a systematic review and Bayesian network meta-analysis of randomised controlled trials. The review was initiated on 1 July 2025 and registered in PROSPERO on 18 July 2025. PubMed, Embase, Web of Science, the Cochrane Library, CNKI, VIP and Wanfang were searched from database inception, defined as the earliest records indexed in each database, to 2 June 2026, together with reference lists and trial registries. Eligible interventions were aerobic exercise (AE), resistance exercise (RE), breathing training (BT), aerobic exercise plus resistance exercise (AE + RE), aerobic exercise plus breathing training (AE + BT), and usual care (UC). Analyses used change-from-baseline values when available and post-intervention endpoint values otherwise. Risk of bias was assessed with RoB 2. Network estimates were reported as mean differences (MDs) with 95% credible intervals (CrIs). Certainty of evidence was assessed using the GRADE approach adapted for network meta-analysis and the Confidence in Network Meta-Analysis (CINeMA) framework.</p> Results <p>Sixteen trials involving 644 participants were included. For health-related quality of life measured using the Short Form-36 (SF-36), AE + RE showed a favourable but imprecise signal, and the credible interval did not exclude no clear effect. For 6-min walk distance (6MWD), BT showed the largest estimated improvement versus UC; however, this finding was based on limited direct evidence and should be interpreted cautiously. AE + BT improved peak oxygen uptake compared with UC. Evidence for forced expiratory volume in 1&#xa0;s percentage predicted (FEV₁%) and N-terminal pro-B-type natriuretic peptide (NT-proBNP) was particularly sparse and imprecise, and no firm conclusions could be drawn for these outcomes. Overall, certainty of evidence was generally low or very low for many active-intervention comparisons because of sparse networks, wide credible intervals, risk of bias and clinical indirectness.</p> Conclusions <p>Supervised exercise-based rehabilitation showed favourable signals for selected outcomes assessed mainly at the end of the intervention in adults with PAH and related pulmonary hypertension populations. However, the evidence remains uncertain because several networks were sparse, credible intervals were wide, direct active-versus-active comparisons were limited, and certainty of evidence was generally low or very low. SUCRA rankings should be interpreted only as exploratory signals and not as confirmatory evidence for choosing one exercise modality over another. In particular, the favourable ranking of AE + RE for SF-36 should not be considered clear superiority for quality of life. Larger, adequately powered and subtype-stratified trials with standardised exercise prescriptions, harmonised safety reporting and longer follow-up are needed.</p> Trial registration <p>PROSPERO CRD420251107670.</p>