Purpose: This study compared the occurrence of different types of generalization (within-class, across-class, and total generalization) following motor-phonetic speech therapy and linguistic-phonological speech therapy in children with a cleft palate ± cleft lip (CP ± L). Method: Thirteen children with a CP ± L (M[subscript age] = 7.50 years) who previously participated in a block-randomized, sham-controlled design comparing motor-phonetic therapy (n = 7) and linguistic-phonological therapy (n = 6) participated in this study. Speech samples consisting of word imitation and sentence imitation were collected on different data points before and after therapy and perceptually assessed using the Dutch translation of the Cleft Audit Protocol for Speech-Augmented. The percentages within-class, across-class, and total generalization were calculated for the different target consonants. Generalization in the two groups was compared over time using linear mixed models (LMMs). Results: LMM revealed significant Time × Group interactions for the percentage within-class generalization in sentence imitation and total generalization in sentence imitation tasks indicating that these percentages were significantly higher in the group of children who received linguistic-phonological intervention. No Time × Group interactions were found for the percentages across-class generalization. Conclusions: Generalization can occur following both motor-phonetic intervention as well as linguistic-phonological intervention. A linguistic-phonological approach, however, was observed to result in larger percentages of within-class and total generalization scores. As children with a CP ± L often receive yearlong intervention to eliminate cleft-related speech sound errors, these findings on the superior generalization effects of linguistic-phonological intervention are important to consider in clinical practice.