Robotic surgery for rectal cancer is gaining popularity, but evidence on long-term oncological outcomes is scarce. We aimed to compare surgical quality and long-term oncological outcomes of robotic and conventional laparoscopic surgery in patients with middle and low rectal cancer. Here we report the short-term outcomes of this trial. This multicentre, randomised, controlled, superiority trial was done at 11 hospitals in eight provinces of China. Eligible patients were aged 18-80 years with middle (>5 to 10 cm from the anal verge) or low (≤5 cm from the anal verge) rectal adenocarcinoma, cT1-T3 N0-N1 or ycT1-T3 Nx, and no evidence of distant metastasis. Central randomisation was done by use of an online system and was stratified according to participating centre, sex, BMI, tumour location, and preoperative chemoradiotherapy. Patients were randomly assigned at a 1:1 ratio to receive robotic or conventional laparoscopic surgery. All surgical procedures complied with the principles of total mesorectal excision or partial mesorectal excision (for tumours located higher in the rectum). Lymph nodes at the origin of the inferior mesenteric artery were dissected. In the robotic group, the excision procedures and dissection of lymph nodes were done by use of robotic techniques. Neither investigators nor patients were masked to the treatment allocation but the assessment of pathological outcomes was masked to the treatment allocation. The primary endpoint was 3-year locoregional recurrence rate, but the data for this endpoint are not yet mature. Secondary short-term endpoints are reported in this article, including two key secondary endpoints: circumferential resection margin positivity and 30-day postoperative complications (Clavien-Dindo classification grade II or higher). The outcomes were analysed according in a modified intention-to-treat population (according to the original assigned groups and excluding patients who did not undergo surgery or no longer met inclusion criteria after randomisation). This trial was registered with ClinicalTrials.gov, number NCT02817126. Study recruitment has completed, and the follow-up is ongoing. Between July 17, 2016, and Dec 21, 2020, 1742 patients were assessed for eligibility. 502 patients were excluded, and 1240 patients were enrolled and randomly assigned to receive either robotic surgery (620 patients) or laparoscopic surgery (620 patients). 69 patients were excluded (34 in the robotic surgery group and 35 in the laparoscopic surgery group). 1171 patients were included in the modified intention-to-treat analysis (586 in the robotic group and 585 in the laparoscopic group). Six patients in the robotic surgery group received laparoscopic surgery and seven patients in the laparoscopic surgery group received robotic surgery. 22 (4·0%) of 547 patients in the robotic group had a positive circumferential resection margin as did 39 (7·2%) of 543 patients in the laparoscopic group (difference -3·2 percentage points [95% CI -6·0 to -0·4]; p=0·023). 95 (16·2%) of patients in the robotic group had at least one postoperative complication (Clavien-Dindo grade II or higher) within 30 days after surgery, as did 135 (23·1%) of 585 patients in the laparoscopic group (difference -6·9 percentage points [-11·4 to -2·3]; p=0·003). More patients in the robotic group had a macroscopic complete resection than in the laparoscopic group (559 [95·4%] of 586 patients vs 537 [91·8%] of 585 patients, difference 3·6 percentage points [0·8 to 6·5]). Patients in the robotic group had better postoperative gastrointestinal recovery, shorter postoperative hospital stay (median 7·0 days [IQR 7·0 to 11·0] vs 8·0 days [7·0 to 12·0], difference -1·0 [95% CI -1·0 to 0·0]; p=0·0001), fewer abdominoperineal resections (99 [16·9%] of 586 patients vs 133 [22·7%] of 585 patients, difference -5·8 percentage points [-10·4 to -1·3]), fewer conversions to open surgery (10 [1·7%] of 586 patients vs 23 [3·9%] of 585 patients, difference -2·2 percentage points [-4·3 to -0·4]; p=0·021), less estimated blood loss (median 40·0 mL [IQR 30·0 to 100·0] vs 50·0 mL [40·0 to 100·0], difference -10·0 [-20·0 to -10·0]; p<0·0001), and fewer intraoperative complications (32 [5·5%] of 586 patients vs 51 [8·7%] of 585 patients; difference -3·3 percentage points [-6·3 to -0·3]; p=0·030) than patients in the laparoscopic group. Secondary short-term outcomes suggest that for middle and low rectal cancer, robotic surgery resulted in better oncological quality of resection than conventional laparoscopic surgery, with less surgical trauma, and better postoperative recovery. Shenkang Hospital Development Center, Shanghai Municipal Health Commission (Shanghai, China), and Zhongshan Hospital Fudan University (Shanghai, China).