PURPOSE: Laboratory studies suggested that caffeine and other nutrients contained in coffee and tea may protect against non-melanoma skin cancer (NMSC). However, epidemiological studies conducted so far have produced conflicting results. METHODS: We performed a literature review and meta-analysis of observational studies published until February 2016 that investigated the association between coffee and tea intake and NMSC risk. We calculated summary relative risk (SRR) and corresponding 95 % confidence intervals (95 % CI) by using random effects with maximum likelihood estimation. RESULTS: Overall, 37,627 NMSC cases from 13 papers were available for analysis. Intake of caffeinated coffee was inversely associated with NMSC risk (SRR for those in the highest vs. lowest category of intake: 0.82, 95 % CI 0.75-0.89, I 2 = 48 %), as well as intake of caffeine (SRR 0.86, 95 % CI 0.80-0.91, I 2 = 48 %). In subgroup analysis, these associations were limited to the basal cell cancer (BCC) histotype. There was no association between intake of decaffeinated coffee (SRR 1.01, 95 % CI 0.85-1.21, I 2 = 0) and tea (0.88, 95 % CI 0.72-1.07, I 2 = 0 %) and NMSC risk. There was no evidence of publication bias affecting the results. The available evidence was not sufficient to draw conclusions on the association between green tea intake and NMSC risk. CONCLUSIONS: Coffee intake appears to exert a moderate protective effect against BCC development, probably through the biological effect of caffeine. However, the observational nature of studies included, subject to bias and confounding, suggests taking with caution these results that should be verified in randomized clinical trials.