PURPOSE: to compare the values of voice onset time (VOT) of plosives, produced by children with phonological disorder and difficulty in the production of voicing contrast, classified as voiceless or voiced from a perceptual-auditory analysis. METHOD: five boys with phonological disorder and difficulty in establishing the feature [+voice] of plosives took part in the study. Based on word pairs (['papa], ['baba], ['tata], ['dada], ['kaka] and ['gaga]) inserted in carrier phrases, the VOT of each plosive was extracted. Based on the same speech sample, three speech therapists carried out a perceptual-auditory analysis, where one should judge the phonemes as voiceless or voiced. The values of VOT as for phonemes classified as voiceless were compared with the values of VOT as for phonemes classified as voiced with the support of statistical tests. RESULTS: we found a statistically significant difference only between the VOT values of [p, d, g] judged as voiceless or voiced in initial onset as well as among the three articulatory points. The other analyzed variables did not show statistical significance. CONCLUSION: in general, the results suggest that VOT is not a decisive cue to the perception of distinction of voicing in cases of phonological disorder. However, it showed influence in the discrimination of phonemes according to articulatory point also in phonological disorder. Regarding the acoustics, we observed that: (a) the presence of pre-voicing influences the judgment of the consonant as voiced; (b) the duration of the positive VOT is not decisive for the distinction between voicing and (c) zero VOT, usually it is responsible for the identification of a voiced plosive.