Traditional meta-analysis offers only direct comparative evidence. The optimal cognitive training for poststroke cognitive impairment (PSCI) remains largely undetermined. This study aims to assess and compare the effectiveness of selected cognitive training methods for PSCI patients and to identify and rank the most effective intervention programs. Searches were conducted in PubMed, Embase, Cochrane Library, Web of science, China National Knowledge Infrastructure, China Science and Technology Journal Database, Wanfang Database, and China Biomedical Database for randomized controlled trials up to September 30, 2023. Two researchers independently performed literature screening, data extraction, and quality assessment. Network meta-analysis was utilized to synthesize the main findings. The primary outcome focused on the intervention's impact on subjective cognitive function, with secondary outcomes including effects on activities of daily living, motor function, and functional independence. This study is registered with PROSPERO (CRD42023463282). Fifty eligible randomized controlled trials were identified, revealing eight distinct interventions. These interventions collectively demonstrate efficacy in enhancing cognition. Traditional cognitive training significantly improves overall cognitive function, daily living function, motor function, and functional independence. In Loewenstein Occupational Therapy Cognitive Assessment, Barthel Index, Fugl-Meyer Assessment, and Functional Independence Measure scales, a combination of computer-based and traditional cognitive training outperformed the conventional control group MD = 29.97 (95%CI: 16.3, 44.2), MD = 18.67 (95%CI: 9.78, 27.45), MD = 28.76 (95%CI: 5.46, 51.79) and MD = 42.2 (95%CI: 5.25, 78.99). In the MMSE scale, virtual reality cognitive training combined with traditional training was most effective MD = 8.01 (95%CI: 3.6, 12.4). On the MoCA scale, the combination of exercise and cognitive training showed superior results MD = 6.68 (95%CI: 2.55, 10.78). Only the combined computer-based and traditional cognitive training, as well as traditional cognitive training alone, significantly enhanced functional independence, with no notable differences in other pairwise interventions. The network meta-analysis suggests that augmenting traditional training with other modalities may enhance overall effectiveness. Specifically, interventions incorporating computer-based cognitive training appear to surpass other methods in improving cognition, daily living function, motor skills, and functional independence. The findings of this network meta-analysis provide evidence-based guidance for clinical decision-making. https://www.crd.york.ac.uk/PROSPERO/, identifier in PROSPERO (CRD42023463282).