The association between famine exposure and type 2 diabetes (T2DM) remains controversial. A meta-analysis was performed to clarify the relation of famine exposure to T2DM risk in detail. A systematic literature search was performed in PubMed, Cochrane library, Web of Science, China National Knowledge Infrastructure (CNKI), Sinomed (CBM), Wanfang Data Knowledge Service Platform and China Science and Technology Journal VIP Database. The articles were limited to be available in English or Chinese before 10 April 2019. The pooled relative risk (RR) with 95% confidence interval (CI) was used to estimate the effect of famine exposure on T2DM. The I square value (I2) was used to assess heterogeneity, and the random effects model (REM) was adopted as the pooling method. This meta-analysis included 24 published articles with 1,660,385 subjects. A significant association of famine exposure with increased risk of T2DM was observed (RR=1.433, 95% CI: 1.101-1.865). The pooled RRs were different in terms of publication year, location, study design, sample size, famine duration and definition of diabetes (criteria). Longer the famine exposure experienced, greater the effect of famine exposure on T2DM would be. Compared to WHO criteria, studies of which diabetes diagnosed by ADA criteria and other criteria might more likely to find this association. Compared to moderate famine exposure, the pooled RR in severe famine was statistically significant. Furthermore, the pooled RR adjusted for the most confounders in females was 1.517 (95% CI: 1.018-2.261). In terms of period of famine exposure, compared to adult exposure, people who exposure famine in fetal, childhood, and adolescent were likely to be diabetes. Famine exposure may increase the risk of T2DM. The risk may vary with famine duration, definition of diabetes, gender and period of famine exposure.