High-intensity interval training (HIIT) and moderate-intensity continuous training (MICT) have been increasingly adopted for localized prostate cancer (PCa) under active surveillance (AS). However, it is unclear which training modality is the most favorable in terms of cardiorespiratory fitness and biochemical progression.We searched PubMed, Cochrane and Embase for relevant RCTs. PRISMA guideline was adopted to ensure optimal conduct of this study. Serum prostate specific antigen (PSA) and peak VO2 were selected as primary outcomes and PSA doubling time (PSADT) and testosterone were selected as secondary outcomes. Only articles written in English were included. Cochrane risk-of-bias tool was used for risk of bias evaluation.A total of 501 studies were selected. Six RCTs with 222 patients were included for data extraction and analysis. High-intensity interval training (HIIT) group demonstrated significantly lower PSA compared with usual care (UC) (MD = -1.4; 95%CI = -2.77 to -0.03) and moderate-intensity continuous training (MICT) group (MD = -1.67; 95%CI = -3.30 to -0.05). Both HIIT and MICT showed significantly improved peak VO2 compared with UC. No significant difference was observed in PSADT and testosterone among different training modalities and UC. Regarding peak VO2, MICT had the highest surface under cumulative ranking curve (SUCRA) scores (98.1%). For serum PSA, HIIT had the highest probability (97.8%) to be the training with the highest efficacy. The potential source of bias mainly came from poorly performed allocation concealment and blinding strategies.The present study indicated that HIIT and MICT showed considerable cardiorespiratory benefits for localized PCa. HIIT was preferred over MICT in biochemical progression control in terms of decreasing serum PSA levels. However, MICT was favored over HIIT regarding cardiorespiratory benefits. The findings of this study may facilitate future lifestyle intervention, particularly in the form of physical training, for individuals diagnosed with localized PCa under AS.© 2024. The Author(s), under exclusive licence to Springer Nature Limited.