Disclosed automated speech recognition system for children with speech impairment comprises a means for vocal tract length normalization, a means for voice transformation and a means for model adaptation for adapting acoustic models onto different feature specifications for aiding speech recognition of children with speech impairment. The means for vocal tract length normalization rescales a frequency axis to compensate for a shorter vocal tract. The automated speech recognition system provides a high performance automated speech recognizer for children by incorporating a plurality of children corpus for helping children to read by correcting their speech. The means for voice transformation normalizes a voice signal using at least one voice transformation technique. The means for model adaptation adapts acoustic characteristics of childrens speech onto different feature specifications for aiding speech recognition of children with speech impairment.