Background: ABO-incompatible liver transplantation (ABOi-LT) is increasingly used to overcome donor shortage. Evidence about disadvantage and advantage in comparison with ABO-compatible liver transplantation (ABOcLT) needs to be performed in the early and late periods. Herein, We compared the short-term and long-term outcomes between ABOi-LT and ABOc-LT cohorts. Methods: We performed a meta-analysis based on the observation studies which included outcomes at >= 1 year after ABOi-LT and ABOc-LT procedures, based on the MEDLINE (via Pubmed), the Cochrance Central Register of Controlled Trials (CENTRAL), and EMBASE (via Ovid) systems. Two researchers independently screened each study according to the pre-established inclusion and exclusion criteria to assess the quality of each study and extracted data from published studies. The primary outcome indicators were all-cause mortality and graft survival at 1, 3 and 5 years after transplantation. In the meta-analysis, we based on the value of heterogeneity using a fixed-effect and a random-effect. A fixed-effect model was used if the value of 12 was less than or equal 50%; and a random-effect model was used if the value of 12 was greater than 50%. Findings: Out of 335 identified records, 29 records with 10,783 patients with liver transplants; 2137 of them were ABOi-LTs and the remaining 8646 were ABOc-LTs. There was no significant difference at 1-year, 3-year, and 5-year in all-cause mortality, death-censored graft survival and complication incidence rate between ABO-incompatible living donor liver transplantation (ABOi-LDLT) group and ABO-compatible living donor liver transplantation (ABOc-LDLT) group. Compared with ABO-compatible deceased donor liver transplantation (ABOc-DDLT), ABO-incompatible deceased donor liver transplantation (ABOi-DDLT) had a higher 1-year all-cause mortality, and the value of totally pooled odds ratio (OR) was 1.89 (1.28,2.80). However, there was no significant difference at 3-year and 5-year all-cause mortality between ABOi-DDLT and ABOc-DDLT groups. ABOi-DDLT group had a lower 1-year and 5-year death-censored graft survival than ABOc-DDLT, as the value of totally pooled OR was 1.91 (1.41,2.60) and 1.52 (1.12,2.05), respectively. No significant difference was detected at 3-year death-censored graft survival between ABOi-DDLT and ABOc-DDLT groups. ABOi-DDLT group had a higher complication incidence rate than ABOc-DDLT, and the value of totally pooled OR was 2.26 (1.53,3.33). We found no obvious bias except for the complication of living donor liver transplantation (LDLT; P = 0.038). In conclusion: The short-term and long-term outcomes were worse after ABOi-DDLT than ABOc-DDLT in the allcause mortality, death-censored graft survival, and complication incidence rate. However, the same outcomes were essentially comparable between ABOi-LDLT vs. ABOc-LDLT cohorts. Considering the current shortage of liver donors, we believe that ABOi-LT from living donor and deceased donors can save lives under emergency situations.