There is currently no detailed evidence for the long-term effects of bariatric surgery on severely obese with type 2 diabetes, such as the risk of myocardial infarction and stroke. In order to provide evidence on the risks of macrovascular diseases and metabolic indicators of bariatric surgery follow-up for more than five years, we searched in the Cochrane library, Pubmed, and EMBASE databases from the earliest studies to January 31, 2019. Randomized clinical trials or cohort studies compared bariatric surgery and conventional medical therapy for long-term incidence of macrovascular events and metabolic outcomes in severely obese patients with T2DM. Fixed-effects and random-effects meta-analyses were performed to pool the relative risks (RRs), hazard ratios (HRs) and weighted mean difference (WMD). Publication bias and heterogeneity were examined. Four RCTs and six cohort studies were finally involved in this review. Patients in the bariatric surgery group as compared to the conventional treatment group had lower incidence of macrovascular complications (RR = 0.43, 95%CI = 0.27~0.70), cardiovascular events (CVEs) (HR = 0.52, 95%CI = 0.39~0.71), and myocardial infarction (MI) (RR = 0.40, 95%CI = 0.26~0.61). At the same time, the results demonstrate that bariatric surgery is associated with better weight and better glycemic control over the long-term than non-surgical therapies, and reveal that different surgical methods have different effects on various metabolic indicators. Bariatric surgery significantly decreases macrovascular complications over the long term and is associated with greater weight loss and better intermediate glucose outcomes among T2DM patients with severe obesity as compared to patients receiving only conservative medical measures.