The national annually reported proportion of men who have sex with men (MSM) among people living with HIV (PLWH) is growing in China. To better inform the public health sector how to improve HIV prevention strategies, it is necessary to understand the current level of HIV incidence and its correlates. Google Scholar, PubMed, Web of Science and three major Chinese electronic publication databases (http://qikan.cqvip.com/, http://g.wanfangdata.com.cn/, http://www.cnki.net/, respectively) were searched for studies reporting HIV incidence. Comprehensive Meta-Analysis (CMA) 2.0 statistical software (Biostat, Inc. Englewood, NJ, USA) was used to calculate the pooled HIV incidence and perform subgroup-analysis to find correlates for HIV seroconversion in Chinese MSM. Random effects modelling was then conducted. Twenty-five eligible studies were included in this meta-analysis. The calculated pooled HIV incidence was 5.61/100 person years (PY), with an increasing trend over time (3.24/100PY, 5.29/100PY, 5.50/100PY in 2005-2008, 2009-2011, 2012-2014 respectively, χ2 test for trend P = 0.04). Subgroup analyses indicated that age <25 years (rate ratio (RR) = 1.85), junior college education and below (RR = 1.87), having ≥ 2 male sexual partners in past 6 months (RR = 2.50), baseline syphilis infection (RR = 2.99), homosexual orientation (RR = 1.91), preferred bottom/versatile roles in anal sexual intercourse (RR = 2.33), and having unprotected anal intercourse in the past 6 months (RR = 2.16) significantly increased the risk for HIV seroconversion (each P < 0.05). Uncircumcised MSM had a marginal statistically significant higher HIV incidence (RR = 3.35, P = 0.051). HIV incidence is still alarmingly high among Chinese MSM. Stronger HIV intervention strategies should be implemented, in particular targeting young, less educated and syphilis-infected MSM.