Purpose: To compare morphosyntactic skills of preschoolers in different subgroups of language impairment. Method: Eighty-three children participated in this study. They represented 4 groups: (a) language impairment-only, (b) speech-language impairment with minimal or no final cluster reduction/consonant deletion, (c) speech-language impairment with frequent final cluster reduction/consonant deletion, and (d) a no-impairment control group. Group performance was compared for finite and nonfinite morpheme production and sentence structure. Results: Children in the language impairment-only group had significantly higher performance than children in both speech-language impairment subgroups, even when errors that could be attributed to final consonant deletion/cluster reduction were taken into account. The language impairment-only and control groups' performance was similar for finite and nonfinite morpheme production, and both groups produced nonfinite plurals with significantly higher accuracy than finite third person singular forms. The language impairment-only group had significantly higher accuracy for both plural and third person singular relative to the group with speech-language impairment characterized by infrequent final cluster reduction/consonant deletion. Conclusions: Children with speech-language impairment generally had poorer morphosyntactic skills than peers who had language deficits and age-appropriate speech skills. Final consonant and final cluster production skills alone did not account for group differences. Clinically, the findings suggest that it is important to assess carefully the speech skills, including final cluster production skills, of preschoolers who have language deficits and language skills of preschoolers who have speech sound disorders. (Contains 4 tables and 1 figure.)