To comprehensively evaluate the therapeutic effects on both functional and structural outcomes, we performed a meta-analysis of preclinical data on stem cell therapy in intracranial hemorrhage, thus providing optimal evidence and instruction for clinical translation. We searched online databases to identify eligible studies based on unmodified stem cell transplantation in intracranial hemorrhage (ICH). From each study, we extracted data regarding neurobehavioral and histological outcomes in order to analyze the comprehensive effective sizes according to the most important clinical parameters (seven indices) and to explore any potential correlation through meta-regression. We analyzed 40 eligible studies including 1021 animals and found a significant improvement in both behavioral and structural outcomes with the median effect size of 1.77 for modified Neurological Severity Score, 1.16 for the modified placement test, 1.82 for the rotarod test, and 1.24 for tissue loss reduction. The meta-regression results revealed that intracerebral administration was the most effective for behavioral and structural recovery post-ICH; mesenchymal stem cells shared comparable therapeutic effects with neural stem cells. Delayed therapy, applied more than 1 week after ICH, showed the greatest improvement of structural outcomes. Stem cell therapy showed significant improvement on behavioral and structural outcomes of ICH animals with relatively large effect sizes. However, the practical efficacy of the therapy is likely to be lower considering poor study quality and non-negligible publication bias. Further, future research should interpret animal results cautiously considering the limited internal and external validity when referring to the design of both animal studies and clinical trials.