Importance Preclinical evidence and case reports suggest potential therapeutic benefits of ketogenic diets (KDs) in the treatment of depression, but evidence from well-controlled randomized clinical trials (RCTs) is lacking. Objective To assess the efficacy of a KD compared with a control diet in adults with treatment-resistant depression (TRD). Design, Setting, and Participants This RCT was conducted between February 22 and June 15, 2024. Participants aged 18 to 65 years with TRD and a score of 15 or greater on the 9-item Patient Health Questionnaire (PHQ-9) from across the UK were included. Intervention Participants were randomized 1:1 to one of two 6-week dietary interventions: (1) KD of prepared foods providing less than 30 g of carbohydrates per day with weekly individual dietetic support or (2) a control (phytochemical [phyto]) diet with vouchers to purchase 1 extra serving of vegetables or fruit and replace saturated fats with unsaturated fats, with equal dietetic support. The last follow-up was at 12 weeks. Main Outcomes and Measures The primary outcome was the between-group difference in change in PHQ-9 score from baseline to week 6. Secondary outcomes included PHQ-9 score at 12 weeks, depression remission, anxiety, anhedonia, cognitive impairment, quality of life, and functional outcomes. Results The study included 88 participants (mean [SD] age, 42.1 [13.1] years; 61 women [69%]): 44 in the KD group and 44 in the phyto diet group. Depression severity decreased markedly in both groups; the mean (SD) change in PHQ-9 score from baseline to week 6 was −10.5 (7.0) in the KD group and −8.3 (5.1) in the phyto group. The mean between-group differences in PHQ-9 score at 6 and 12 weeks were −2.18 (95% CI, −4.33 to −0.03; P =.05; Cohen d, −0.68; 95% CI −1.35 to −0.01) and −1.85 (95% CI, −4.04 to 0.33; P =.10; Cohen d, −0.58; 95% CI, −1.26 to 0.10), respectively. There were no differences in secondary outcomes between the KD and phyto groups. No serious adverse events occurred. Conclusions and Relevance In this RCT, a KD had antidepressant benefits compared with a well-matched control diet at 6 weeks. However, the clinical relevance is uncertain, as the mean effect size compared with the control was modest and not evident in secondary analyses.