OBJECTIVE For large colorectal tumors, the en bloc resection rate achieved by endoscopic mucosal resection (EMR) is insufficient, and this leads to a high rate of local recurrence. As endoscopic submucosal dissection (ESD) has been reported to achieve a higher rate of en bloc resection and a lower rate of local recurrence in the short-term, it is expected to overcome the limitations of EMR. We conducted a matched case-control study between ESD and EMR to clarify the effectiveness of ESD for colorectal tumors. METHODS Between April 2005 and February 2009, a total of 28 colorectal tumors in 28 patients were resected by ESD and were followed up by colonoscopy at least once. As a control group, 56 EMR cases from our prospectively completed database were matched. En bloc resection, complication and recurrence rates were compared between the two groups. RESULTS The mean sizes of the lesions were 27.1 mm in the ESD group and 25.0 mm in the EMR group. The en bloc resection rate was significantly higher in the ESD group (92.9% vs 37.5% with ESD vs EMR), and the rate of perforation was also significantly higher (10.7% vs 0%). All cases of perforation were managed conservatively. No recurrence was observed in the ESD group, whereas local recurrences were detected in 12 EMR cases (21.4%). Eleven of the 12 recurrences (91.7%) were managed endoscopically, and one required surgical resection. CONCLUSIONS Endoscopic submucosal dissection is a promising technique for the treatment of colorectal tumors, giving an excellent outcome in comparison with EMR.