Children with High-Functioning Autism's (HFA) sensitivity to emotional prosody in speech sounds, was long under debate. The stiffness prohibiting them to switch between diverse features might be one of the factors that make them difficult to understand emotional prosody in daily communications. Present study explored how dimension-switch ability in identifying lexical tone would modulate children's perception of emotional prosody, since both prosody and lexical tone are conveyed by pitch height and contour. Mandarin-speaking HFA (n = 22) and typically developing (TD, n = 16) children aged 7 to 12 first classified emotions of 24 emotion-prosody congruent/incongruent words. Second, they identified lexical tone in a categorical perception task, and then received a dimension switch task, which asked them to choose the wrong answer. Two groups of children performed similarly in these tasks. HFA children identified basic emotions well as TD children even under the incongruent word pairs, which could be explained by their even performances in dimension switch task. Despite the similarity, there was a correlation between the accuracy rate of emotion classification and lexical tone perception in HFA children, which was not seen in TD children. This showed that HFA children used correlated mechanisms in understanding emotional prosody and lexical tone.