IMPORTANCE: Evidence from everyday dining situations regarding the effects of traffic light labels (TLLs) on dietary improvement remains inconsistent. OBJECTIVE: To evaluate the effects of TLLs on dietary consumption and choices in cafeteria settings. DESIGN, SETTING, AND PARTICIPANTS: This 2-arm, parallel randomized clinical trial was conducted at a company staff cafeteria in Shanghai, China, from September to December 2022. Of 153 adult participants, 76 were randomly assigned to the intervention group and 77 to the control group. Data analysis was conducted from July to October 2024. INTERVENTION: The intervention group gained access to TLLs offering a comprehensive rating of added sugar, fat, and sodium for each dish on the lunch menu, while the control group did not. MAIN OUTCOMES AND MEASURES: Primary outcomes were lunch intake of added sugar, fat, and sodium. Secondary outcomes included the mean traffic light score (calculated based on the number of dishes consumed, with higher scores indicating worse overall dietary choices) and number of green-coded (reaching dietary recommendations), yellow-coded (between the recommendation and mean intake of the Chinese population), and red-coded (above the upper limit of intake) dishes. The primary and secondary outcomes were automatically calculated based on the precollected recipe dataset of the cafeteria and the Chinese food composition database when participants ordered meals using an applet during weekday lunchtime. Weekly median values of these outcomes were used for analyses, which were conducted for the intention-to-treat population. RESULTS: Among 153 participants, the mean (SD) age was 32.7 (7.5) years, and 97 (63.4%) were female. At week 12, compared with the control group, the intervention group demonstrated no statistically significant decrease in dietary consumption of added sugar (mean difference, -0.15 [95% CI, -0.75 to 0.46] g), fat (mean difference, -1.54 [95% CI, -6.13 to 3.05] g), or sodium (mean difference, -116.12 [95% CI, -454.78 to 222.54] mg). Similarly, no statistically significant differences were observed in dietary choices based on the mean traffic light score (mean difference, -0.05 [95% CI, -0.12 to 0.03]) or the number of green-coded (odds ratio [OR], 1.15 [95% CI, 0.99-1.32]), yellow-coded (OR, 1.04 [95% CI, 0.90-1.20]), and red-coded (OR, 0.84 [95% CI, 0.57-1.23]) dishes. CONCLUSIONS AND RELEVANCE: In this randomized clinical trial, TLLs indicating added sugar, fat, and sodium ratings on menus failed to improve dietary consumption and choices in a company cafeteria setting. This finding suggests that TLLs on menus may not effectively promote dietary improvement in this setting. TRIAL REGISTRATION: Chinese Clinical Trial Registry Identifier: ChiCTR2100051771.