The aim of the present study is to clarify the relationship between nasalance scores and nasopharyngeal shapes obtained by lateral cephalograms.Eight patients who underwent a Wardill–Kilner push-back palatoplasty were included in this study. Perceptual judgment by a speech pathologist indicated that these patients had no hypernasality and no nasal emission at blowing. As normal controls, 33 non-cleft individuals, 4 boys and 10 girls aged 6 years old and 5 boys and 14 girls aged 7 years old, were investigated.Lateral cephalograms at rest were taken for both groups. For the cleft (palate) patients, lateral cephalograms at phonation /a/ and blowing were analyzed and nasometries were also performed using a kitsutsuki passage.There was no significant difference in the velar length, the pharyngeal depth, the ratio of the velar length to the pharyngeal depth and the velar angle between the cleft patients and the non-cleft individuals. Multiple regression analyses indicated that standardized regression coefficients of ratios for the velar length to the pharyngeal depth and the velar ascent at blowing had higher nasalance scores for sentences 1 and 3, which had high coefficients of determination, respectively.