Advances in virtual endoscopy simulators have paralleled an interest in medical simulation for gastrointestinal endoscopy training. The primary objective was to determine whether the virtual endoscopy simulator training could improve the performance of novices. A systematic review. Randomized controlled trials (RCTs) that compared virtual endoscopy simulator training with bedside teaching or any other intervention for novices were collected. Novice endoscopists. The PRISMA statement was followed during the course of the research. The Cochrane Central Register of Controlled Trials, MEDLINE, EMBASE, and ScienceDirect were searched (up to July 2013). Data extraction and assessment were independently performed. Independent procedure completion, total procedure time and required assistance. Fifteen studies (n = 354) were eligible for inclusion: 9 studies designed for colonoscopy training, 6 for gastroscopy training. For gastroscopy training, procedure completed independently was reported in 87.7% of participants in simulator training group compared to 70.0% of participants in control group (1 study; 22 participants; RR 1.25; 95% CI 1.13-1.39; P<0.0001). For colonoscopy training, procedure completed independently was reported in 89.3% of participants in simulator training group compared to 88.9% of participants in control group (7 study; 163 participants; RR 1.10; 95% CI 0.88-1.37; P = 0.41; I(2) = 85%). The included studies are quite in-homogeneous with respect to training schedule and procedure. Virtual endoscopy simulator training might be effective for gastroscopy, but so far no data is available to support this for colonoscopy.