We attempted to establish clinical criteria which differentiated between recessive dystrophic epidermolysis bullosa (RDEB) and dominant DEB (DDEB), since these two groups show prominent differences in prognosis, genetic recurrence risk and response to some types of treatment. The total of 119 cases examined consisted of our own 9 cases (2 of RDEB and 7 of DDEB) and 110 cases (26 of RDEB and 84 of DDEB) collected from the medical literature. They were analyzed by calculating the sensitivity and specificity of ten clinical features. We concluded that, syndactyly, presence of complications, dental lesions, remission-less course, and oral lesions are strongly indicative of RDEB (more than 70% specific). Intractable skin ulcer is suggestive of RDEB (more than 55% specific). Nail lesions, scar and atrophy, milia and pigmentation are not helpful in the differentiation because of their commonality. The proposed criteria are simple, reliable, and practical, providing us with a useful tool for differentiation of RDEB and DDEB in daily practice.