Background: The efficacy of exercise interventions in the treatment of mental health disorders is well known, but research is lacking on the most efficient exercise type for specific mental health disorders. Objective: The present study aimed to compare and rank the effectiveness of various exercise types in the treatment of mental health disorders. Methods: The PubMed, Web of Science, PsycINFO, SPORTDiscus, CINAHL databases, and the Cochrane Central Register of Controlled Trials as well as Google Scholar were searched up to December 2021. We performed pairwise and network meta-analyses as well as meta-regression analyses for mental health disorders in general and each type of mental health disorder, with alterations in symptom severity as the primary outcome. Results: A total of 6456 participants from 117 randomized controlled trials were surveyed. The multimodal exercise (71%) had the highest probability of being the most efficient exercise for relieving depressive symptoms. While resistance exercise (60%) was more likely to be the most effective treatment for anxiety disorder, patients with post-traumatic stress disorder (PTSD) benefited more from mind–body exercise (52%). Furthermore, resistance exercise (31%) and multimodal exercise (37%) had more beneficial effects in the treatment of the positive and negative symptoms of schizophrenia, respectively. The length of intervention and exercise frequency independently moderated the effects of mind–body exercise on depressive (coefficient = 0.14, p =.03) and negative schizophrenia (coefficient = 0.96, p =.04) symptoms. Conclusion: Multimodal exercise ranked best for treating depressive and negative schizophrenic symptoms, while resistance exercise seemed to be more beneficial for those with anxiety-related and positive schizophrenic symptoms. Mind–body exercise was recommended as the most promising exercise type in the treatment of PTSD. However, the findings should be treated with caution due to potential risk of bias in at least one dimension of assessment and low-to-moderate certainty of evidence. Trial Registration This systematic review was registered in the PROSPERO international prospective register of systematic reviews (CRD42022310237).