Asperger's syndrome (AS) is a pervasive developmental disorder that is associated with marked social dysfunction. Deficits in the perception of nonverbal cues of emotion may be related to this social impairment. Research has indicated that children with autism are limited in their emotion perception abilities, but studies that have addressed this issue with individuals with AS or high-functioning autism (HFA) have yielded inconsistent findings. These inconsistencies may be related to methodological differences across studies including diagnostic criteria and failure to consider the intensity of the emotion cues. It was hypothesized that children with AS and HFA would both have deficits in emotion perception compared to typically-developing children. However, children with HFA were expected to have an even greater emotion perception deficit than children with AS and this difference was hypothesized to be most pronounced for low intensity cues of emotion. It is important to clarify whether individuals with AS and HFA differ in emotion perception because most studies of this skill combine them into one group or use poorly defined diagnostic criteria. This study examined the ability of 30 8- to 15-year-old children with either AS or HFA to perceive emotion from high and low intensity cues. In order to address limitations with the differential validity of the DSM-IV criteria for AS, diagnoses were based on diagnostic criteria proposed by Klin et al. (in press). A researcher who was blind to diagnosis administered a test that presented low and high intensity cues of emotion in photographs of facial expression and audiotapes of tone of voice. Comparison of the emotion perception accuracy of children with AS to the normative means of this instrument for typically-developing children did not reveal any significant differences. In contrast, the children with HFA were significantly less accurate in their perception of facial expressions and tone of voice than the normative sample and the participants with AS. Contrary to expectations, IQ was significantly related to emotion perception accuracy. After controlling for IQ, the difference in perception of facial expressions between children with AS and HFA was not significant. On the other hand, cue intensity moderated the relation between diagnosis and emotion perception accuracy for tone of voice even after IQ was taken into account. Children with AS perceived high and low intensity tone of voice cues with similar accuracy, but children with HFA had significantly poorer performance on the low intensity tone of voice cues. Although emotion perception accuracy was related to better adjustment, it was not correlated with the most sensitive measure of current social functioning. This suggests that even when children with AS or HFA perceive cues correctly, they may not know how or be able to properly integrate them for adaptive responses in social interaction. The findings have important implications for understanding inconsistencies in past research and identifying future directions.