The goal of our work is to study the reflection of the emotional state in the characteristics of voice and speech of children with Down syndrome (DS). Participants of the study were 20 children aged 8-12 years: 10 children with DS and 10 typically developing (TD) children. Three types of experimental methods of speech analysis were performed: perceptual (by listeners, n = 350), spectrographic, and linguistic. The results of perceptual experiment showed that the listeners better recognized the emotional state of children with DS than TD children, the state of discomfort in children with DS than the neutral and comfort state. Acoustic features reflecting emotional states are more expressed in speech of children with DS vs. speech of TD children. The state of discomfort in the speech of children with DS is described by a large number of acoustic features that differs significantly from the corresponding characteristics reflecting neutral and comfort states. We presented the first data on the emotional speech of children with DS on the material of the Russian language. These data may be in further used for creating human-computer interfaces for informants with atypical development.