OBJECTIVE oral narrative is a means of language development assessment. However, standardized data for deaf patients are scarce. The aim here was to compare the use of narrative competence between hearing-impaired and normal-hearing children. METHODS analytical cross-sectional study at the Department of Speech-Language and Hearing Sciences, Universidade Federal de São Paulo. METHODS twenty-one moderately to profoundly bilaterally hearing-impaired children (cases) and 21 normal-hearing children without language abnormalities (controls), matched according to sex, age, schooling level and school type, were studied. A board showing pictures in a temporally logical sequence was presented to each child, to elicit a narrative, and the child's performance relating to narrative structure and cohesion was measured. The frequencies of variables, their associations (Mann-Whitney test) and their 95% confidence intervals was analyzed. RESULTS the deaf subjects showed poorer performance regarding narrative structure, use of connectives, cohesion measurements and general punctuation (P < 0.05). There were no differences in the number of propositions elaborated or in referent specification between the two groups. The deaf children produced a higher proportion of orientation-related propositions (P = 0.001) and lower proportions of propositions relating to complicating actions (P = 0.015) and character reactions (P = 0.005). CONCLUSIONS hearing-impaired children have abnormalities in different aspects of language, involving form, content and use, in relation to their normal-hearing peers. Narrative competence was also associated with the children's ages and the school type.