Added-sugar consumption in the U.S. exceeds recommended limits. Policymakers are considering requiring restaurants to use menu warning labels to indicate items high in added sugar. We sought to determine whether icon-only and icon-plus-text added-sugar menu labels were (1) perceived as more effective at potentially reducing consumption of items high in added sugar and (2) increased knowledge of menu items' added-sugar content relative to control labels, and if effects differed by label design. A national sample of U.S. adults (n = 1327) participated in an online randomized experiment. Participants viewed menu items with either a control label, 1 of 6 icon-only labels, or 1 of 18 icon-plus-text labels with 3 text variations. For their assigned label, participants provided ratings of perceived message effectiveness (a validated scale of a message's potential to change behavior). Participants were also asked to classify menu items by their added-sugar content. The icon-only and icon-plus-text labels were perceived as more effective than the control label (means: 3.7 and 3.7 vs. 3.1, respectively, on a 5-point scale; p < 0.001). The icon-only and icon-plus-text groups each correctly classified 71% of menu items by added-sugar content vs. 56% in the control group (p < 0.001). All icons and text variations were perceived as similarly effective. In conclusion, relative to a control label, icon-only and icon-plus-text added-sugar menu labels were perceived as effective and helped consumers identify items high in added sugar. Menu warning labels may be a promising strategy for reducing added-sugar consumption from restaurants, but research on behavioral effects in real-world settings is needed. Clinical Trials Identifier:NCT04637412.