Reducing harmful consumption of food (including non-alcoholic drinks) and alcoholic drinks would prevent much disease. Evidence from tobacco control shows that placing health warning labels (HWLs) on these products reduces their selection and consumption. A systematic review with meta-analysis was conducted using Cochrane methods to assess the impact on selection (including hypothetical selection) or consumption of food or alcoholic drinks with image-and-text (sometimes termed ‘pictorial’) and text-only HWLs placed on these products. Studies were required to be randomised or quasi-randomised controlled trials. Fourteen studies were included, three on alcohol, eleven on food. For our primary outcomes, eleven studies measured selection and one measured consumption (two measured only other secondary outcomes). Meta-analysis of twelve comparisons from nine studies (n=12,635) found HWLs reduced selection of the targeted product compared to a control group with no HWL (RR=0.74 (95% CI 0.68 to 0.80)), with participants 26% less likely to choose a product displaying a HWL. A planned subgroup analysis suggested a larger effect on selection of image-and-text HWLs (RR=0.65 (95% CI 0.54 to 0.80)) than text-only HWLs (RR=0.79 (95% CI 0.74 to 0.85)), but this difference was not statistically significant. These findings suggest significant potential for HWLs to reduce selection of food and alcohol, but all experimental studies to date were conducted in laboratory or online settings with outcomes assessed immediately after a single exposure. Studies in field and more naturalistic laboratory settings are urgently needed to estimate the potential effects of HWLs on food and alcohol.