The purposes of the present study were the following: 1) to determine the efficacy of our therapy method by comparing the participant's pre- and post-scores of the Japanese Articulation Test (Revised), 2) to compare his overall speech intelligibility via his conversational speech between pre- and post-therapy periods to investigate whether our therapy generalize his accurate pronunciation in a word level to his conversational level, and 3) which kind of the prompt was the most effective to enhance the participants' speech intelligibility. A seventh grade student with Down syndrome was participated in this study. We used the picture flash cards as our teaching material, and conducted instructions to enhance his accurate productions of /s/ and/dz/ phonemes using three kinds of prompts: 1) showing finger signs, 2) Japanese characters (Hiragana) as visual stimuli, and 3) presenting correct sounds as auditory stimuli to remind the participant to speak clearly. These prompts were presented only when he did not pronounce the target phonemes correctly. The results showed that 1) the participant's intelligibility of the target phonemes in a word level was improved, 2) the most effective type of the prompt was showing finger signs, followed by presenting correct sounds, and showing Japanese characters, 3) the scores of the Japanese Articulation Test (Revised) and the Test of Phonological Processes did not change significantly in terms of their pre- and post-scores; however, after the therapy period (B), his error patterns became more constant and started to produce sounds which were close to accurate, and 4) as the participant's speech intelligibility was improved in a word level, his speech intelligibility in a conversational level was also improved. The implications of these results are discussed.