Purpose: To compare and analyze the effects of various Tai Chi styles as an intervention on type 2 diabetes mellitus(T2DM) considering the intensity, on which different opinions have been given. Methods: We searched articles from 12 databases in December 2014. We selected randomized controlled clinical trial (RCT) using Tai Chi on adult patients with T2DM. We performed quality assessment using Cochrane risk of bias (RoB) tool. And we conducted a meta-analysis according to the styles of Tai Chi. Results: We found 13 RCTs within the inclusion criteria. 8 RCTs were studies about Yang style Tai Chi, 4 RCTs were about ‘Tai Chi for diabetes’ and 1 RCT was about Chen style Tai Chi. Studies about Yang and Chen styles of Tai Chi showed significant effect on T2DM, but studies about ‘Tai Chi for diabetes’ didn’t. The meta-analysis showed positive results of ‘Yang style Tai Chi’ for fasting blood glucose (FBG) compared to routine treatment, but failed to show the effects on Hemoglobin A1c (HbA1c). The meta-analysis of ‘Tai Chi for diabetes’ studies showed unfavorable effects on HbA1c compared to routine treatment. And the meta-analysis of 6 studies about all styles of Tai Chi showed unfavorable effects on HbA1c compared to routine treatment.